an open source approach to social media data gathering
TRANSCRIPT
Air Force Institute of TechnologyAFIT Scholar
Theses and Dissertations Student Graduate Works
3-22-2018
An Open Source Approach to Social Media DataGatheringAnthony J. Kallhoff
Follow this and additional works at: https://scholar.afit.edu/etd
Part of the Operational Research Commons
This Thesis is brought to you for free and open access by the Student Graduate Works at AFIT Scholar. It has been accepted for inclusion in Theses andDissertations by an authorized administrator of AFIT Scholar. For more information, please contact [email protected].
Recommended CitationKallhoff, Anthony J., "An Open Source Approach to Social Media Data Gathering" (2018). Theses and Dissertations. 1935.https://scholar.afit.edu/etd/1935
AN OPEN SOURCE APPROACH TO SOCIAL MEDIA DATA GATHERING
THESIS
Anthony J. Kallhoff, 2nd Lieutenant, USAF
AFIT-ENS-MS-18-M-130
DEPARTMENT OF THE AIR FORCE AIR UNIVERSITY
AIR FORCE INSTITUTE OF TECHNOLOGY
Wright-Patterson Air Force Base, Ohio
DISTRIBUTION STATEMENT A.
APPROVED FOR PUBLIC RELEASE; DISTRIBUTION UNLIMITED.
The views expressed in this thesis are those of the author and do not reflect the official
policy or position of the United States Air Force, Department of Defense, or the United
States Government. This material is declared a work of the U.S. Government and is not
subject to copyright protection in the United States.
AFIT-ENS-MS-18-M-130
AN OPEN SOURCE APPROACH TO SOCIAL MEDIA DATA GATHERING
THESIS
Presented to the Faculty
Department of Operational Sciences
Graduate School of Engineering and Management
Air Force Institute of Technology
Air University
Air Education and Training Command
In Partial Fulfillment of the Requirements for the
Degree of Master of Science in Operations Research
Anthony J. Kallhoff, BS
2nd Lieutenant, USAF
March 2018
DISTRIBUTION STATEMENT A.
APPROVED FOR PUBLIC RELEASE; DISTRIBUTION UNLIMITED.
AFIT-ENS-MS-18-M-130
AN OPEN SOURCE APPROACH TO SOCIAL MEDIA DATA GATHERING
Anthony J. Kallhoff, BS
2nd Lieutenant, USAF
Committee Membership:
Dr. Bradley Boehmke
Chair
Maj. Jason Freels
Reader
iv
AFIT-ENS-MS-18-M-130
Abstract
Modern usage of social media affords the military intelligence and analytic
communities novel approaches to gather information. However, the tools and resources to
develop these methodologies are still maturing. Furthermore, current data acquisition
tools are not available to the DoD for all social media platforms. This thesis addresses a
small subset of this problem by developing an open source methodological approach to
collect and manage data from a popular social media site that has previously been
inaccessible to defense intelligence organizations. This approach was operationalized via
the R package called instaExtract, and an exemplar analysis was performed to
demonstrate its application and efficiency for intelligence gathering.
v
Acknowledgments
I would like to express my sincere appreciation to my faculty advisor, Dr. Bradley
Boehmke, for his guidance and support throughout the course of this thesis effort. The
insight and experience was certainly appreciated. I would also like to thank my sponsor
for both the support and latitude provided to me in this endeavor.
Anthony J. Kallhoff
vi
Table of Contents
Page
Abstract .............................................................................................................................. iv
Acknowledgments................................................................................................................v
Table of Contents ............................................................................................................... vi
List of Figures .................................................................................................................. viii
List of Tables ..................................................................................................................... ix
I. Introduction .....................................................................................................................1
1.1 General Issue ..........................................................................................................1
1.2 Research Goals .......................................................................................................2
1.3 Research Contributions ..........................................................................................2
1.4 Assumptions and Limitations .................................................................................3
1.5 Organization ...........................................................................................................3
II. Literature Review ............................................................................................................4
2.1 Overview ................................................................................................................4
2.2 The Evolving State of Intelligence Gathering ........................................................4
2.3 Instagram ................................................................................................................6
2.4 Data Collection and Management ........................................................................10
2.5 Existing Instagram Data Extraction Software ......................................................14
2.6 Reproducible Analysis..........................................................................................18
III. Methodology ...............................................................................................................19
3.1 Overview ..............................................................................................................19
3.2 A JSON-Based Instagram Scrapper in R– instaExtract ................................19
3.3 Adaptable, Reproducible, Distributable – Hosting on the AFIT Data Science
Lab’s GitHub Page .....................................................................................................29
vii
3.4 Summary...............................................................................................................37
IV. Exemplar Analysis .......................................................................................................38
4.1 Overview ..............................................................................................................38
4.2 Gathering Information on the Region...................................................................38
4.3 Hashtag Investigation ...........................................................................................39
4.4 User Investigation .................................................................................................44
4.5 Geo Mapping ........................................................................................................49
4.6 Summary...............................................................................................................51
V. Conclusions and Recommendations ............................................................................52
5.1 Overview ..............................................................................................................52
5.2 The instaExtract Package ..............................................................................52
5.3 The AFIT Data Science Lab .................................................................................54
5.4 Summary...............................................................................................................55
Appendix A ........................................................................................................................56
Appendix B ........................................................................................................................57
Appendix C ........................................................................................................................58
Appendix D ........................................................................................................................59
Appendix E ........................................................................................................................60
Appendix F.........................................................................................................................61
Bibliography ......................................................................................................................62
viii
List of Figures
Page
Figure 1. An Instagram Post ............................................................................................... 8
Figure 2. An Instagram Account ......................................................................................... 9
Figure 3. An Instagram Location ...................................................................................... 10
Figure 4. The Analytic Cycle (Wickham & Grolemund, 2016) ....................................... 11
Figure 5. Word Cloud for Top 200 Hashtags in the DC Area .......................................... 41
Figure 6. Count of 150 Most Used Hashtags in the DC Area........................................... 42
Figure 7. Most liked Instagram Post by usairforce (usairforce, 2018) ............................. 45
Figure 8. Post Likes Over Time for usairforce Account ................................................... 46
Figure 9. Network of Commenting Users on usairforce Posts Where Edge is Proportional
to Number of Comments ............................................................................................ 48
Figure 10. All Washington DC Locations from Instagram Overlaid with Actual Location
.................................................................................................................................... 49
Figure 11. Clustered Locations for Washington DC ........................................................ 50
Figure 12. Zoomed in View of White House Instagram Locations .................................. 51
ix
List of Tables
Page
Table 1 - Instagram Data Extraction Software ................................................................. 15
Table 2 - instaR Functions ................................................................................................ 16
Table 3 - Instagram-php-scraper Functions ...................................................................... 17
Table 4 - Search Functions Provided by the instaExtract Package ........................... 22
Table 5 - Get Functions for the instaExtract Package ............................................... 23
Table 6 - Format of the Results for the Get Functions in the instaExtract Package .. 24
Table 7 - Functions Enabling Location Mapping Capabilities ......................................... 33
1
AN OPEN SOURCE APPROACH TO SOCIAL MEDIA DATA GATHERING
I. Introduction
1.1 General Issue
The modern world is producing mountains of information that can be of
indispensable benefit to the Department of Defense (DoD) intelligence community.
Traditionally, intelligence has been gathered through arduous means and required
advanced systems or highly trained professionals. However, the emergence of social
media, and the breadth of new information it provides, offer analysts a unique
opportunity to access large amounts of data at little cost. While the techniques to make
the most of this data are still in development, they present an exciting supplement to
traditional intelligence gathering.
In academic and applied research, there is an emphasis on advancing and applying
rigorous analytical techniques to data, but also important are the methods to acquire the
data. Just as higher resolution microscopes provide new insights in biology, tools that can
offer new and refined streams of information can help to better inform analysts. Without
attention being given to the acquisition of data, analysts are forced to apply their
techniques to flawed or constrained data sets. Current data gathering tools have varying
limitation in their scope, functionality, and platform, restricting the benefits they may
offer.
2
1.2 Research Goals
The current suite of in-house data acquisition tools does not span all the various
social media platforms. One notable site that is presently excluded is Instagram, a photo
and video sharing platform. While there exist third-party programs that offer some form
of information gathering from Instagram, they are limited in varying ways. This research
aims to establish a reliable and reproducible approach to extract Instagram data for use by
the DoD intelligence community. To this end, this research seeks to create an R-based
Instagram data acquisition package with the following functionality:
• Information retrieval without the need for user authorization or API integration
• Search functions for identifying users and hashtags in Instagram’s databases
• Media post retrieval that includes meta-data and information on community
involvement
• Retrieval of information for users and their activity on the site
• Raw data stream cleaning and transformation resulting in user friendly data
• Example analytic functions to serve as a basis for future development
• Hosting on a platform that facilitates community involvement, continued support,
and version control
1.3 Research Contributions
This research aims to expand the DoD intelligence community’s information
gathering capabilities with a consistent and robust approach to pulling Instagram data.
This approach will be in the form of an R package hosted on the Air Force Institute of
Technology (AFIT) Data Science Lab’s GitHub account, consistent with the approach of
recent analytic tools developed by AFIT. More than a one-off deliverable, the results of
3
this work will establish a groundwork for more Instagram based intelligence gathering
and analysis. By focusing on the particulars of data acquisition, this research supports the
efforts of data analytics by providing the material to investigate. This package will also
have an immediate impact on the DoD, who will have access to the information and
insights Instagram provides.
1.4 Assumptions and Limitations
In development of these resources, some level of continued success by Instagram
and the usefulness of its information is assumed. The tools developed also rely on a
general structure to how Instagram presents its data and allows users to interact with that
data. Significant reworkings to Instagram’s procedures might invalidate some functions
in this research.
1.5 Organization
The paper begins with a literature review in Chapter 2 that explores intelligence
gathering and its evolution with social media. Furthermore, Chapter 2 looks at the current
resources for collection social media data, diving specifically into Instagram, its layout
and tools that interact with it. Finally, Chapter 2 addresses the topic of reproducibility in
analysis. Chapter 3 discusses the methodological framework for extracting Instagram
social media data and the open source R package, instaExtract, that operationalizes
this approach. Next, Chapter 4 presents an exemplar analysis, showing how an analysis
might utilize instaExtract. Finally, Chapter 5 summarizes the work accomplished
with this thesis and discusses areas for future research.
4
II. Literature Review
2.1 Overview
The objective of this thesis is to develop and operationalize a methodological
approach to collect and manage data from a popular social media site as well as provide
an R package that allows for reproducible analysis by users. To this end, this paper
begins with a literature review broken into six sections, including the overview. The
second section will be a look at the more traditional fields of intelligence gathering and
how computers are being integrated with older methods. The third section will look
closer at the specific social media platform explored in this thesis, Instagram, and the
structure of its information. The fourth section covers practices for collecting and
managing large amounts of user data pulled from a social media site using its application
program interface (API). Fifth, the review examines existing data extraction software and
their advantages and limitations. The final section explores techniques and considerations
to ensure reproducible analysis.
2.2 The Evolving State of Intelligence Gathering
The gathering of intelligence has developed throughout the history of warfare.
Knowledge of the enemy and of the landscape of the battle can be vital to the success of
the combatants. Accordingly, the U.S. has developed many institutions and practices for
the purpose of acquiring and deciphering intelligence (Richelson, 2015). Many of the
techniques relied on, however, are fallible and inconsistent. Combing through traditional
paperwork can take many man hours, and even experienced eyes can miss important
details. Many human intelligence procedures also depend on investigators trusting their
memories or interpreting ambiguous signs (Loftus, 2011). With the advent of computers,
5
the nature of intelligence gathering is shifting rapidly. Digitized files can be searched in
milliseconds, and while computers carry their own sets of faults, they do not tire or relent.
More than just work horses though, computers offer unique opportunities for data
acquisition and analysis (Leese, 2015). With the internet came the introduction and
spread of social media sites like Facebook and Instagram. These sites are populated with
information from users all over the world and represent almost all demographics. There is
such a wealth of information that has sprung up so recently, that the intelligence
community is struggling to adapt fast enough to utilize all the data now at its disposal
(Loftus, 2011). Along with a flood of new information, computers are powerful
processers of this information. As an example, consider the analysis of terrorist
networks. Developing an organizational structure of terrorist’s cells has been a backbone
of human intelligence in the fight against terror. Isolating important targets and
determining the flow of information has allowed us to neutralize cells with minimal
damage (Ronczkowski, 2011). However, these models are currently still heavily reliant
on the same unreliable techniques mentioned early. Computers and the new sources of
information they provide could afford the intelligence communities new manners of
network analysis and formulation. Specifically, social media can provide information
unlike any other source. Social media data provides relational and geospatial data at a
real time pace, closing the turnaround of reports for analysts (Pitic, Volovici, Tara, &
Mite, 2013).
The structure of social media sites and the way users interact with them provide
new types of information to be used in analysis. One of the primary and most useful of
these types comes in the form of social networks, a network of actor nodes and
6
relationship or interaction edges that bind them together (Aggarwal, 2011). While social
networks have been studied in conventional scenarios for some time, social media data
presents itself in a manner so conducive to social network analysis that it has created a
renewed fervor in the scientific study of the field (Aggarwal, 2011). These online social
networks tend to reach large scales which manifest with special structures and behaviors
(Milgram, 1967). Online networks in particular are saturated with data, offering unique
circumstances for analysis of two primary kinds of data: structural analysis which
identifies important nodes, links and regions, and content-based analysis which delves
into the specifics of the content being shared (Aggarwal, 2011). Finally, online networks
evolve in much shorter time spans than conventional networks, a complication that is the
subject of recent research (Aggarwal & Yu, 2005).
The previous paragraph mentions content-based analysis briefly, but the scope of
analysis to be done on social media data is extensive. Almost all social media platforms
are rich in text based information which lends itself to interesting fields of analysis like
classification, clustering, and sentiment analysis (Aggarwal, 2011; Ashraf, Verma, &
Tech, 2016). Often, social media data will also contain image and video content, enabling
methods involving image and voice recognition (Li, Zha, Huet, & Tian, 2016). One other
category of information, geotagging, is a descriptor of content, but can provide for
meaningful analysis in its own right. Location information can help to contextualize
online content in the real world (Andriopoulou & Lymberopoulos, 2012).
2.3 Instagram
Instagram launched on 6 October 2010 with 25,000 sign-ups on the first day
(“Our Story – Instagram,” n.d.). Two years later, on 9 April 2012, the company, now at
7
30 million users, was purchased by Facebook for nearly $1 billion (Chaykowski, 2016).
Today, it boasts 500 million daily users, and more than 800 million monthly active users
(“Our Story – Instagram,” n.d.). Instagram is, at its core, just a photo-sharing app, but
with more users than Twitter, Snapchat, and Pinterest combined, it has placed itself as
one of the fastest growing social media platforms of all time (Chaykowski, 2016).
The typical Instagram experience revolves around two types of content: a post and
a story. A post, seen in Figure 4, is the term for a user submission encompassing their
submitted media, description, and community response. A post can either be a video or a
photo. It is accompanied with a user created caption and hashtags, markers that promote
and categorize a post. Posts can also have comments and likes left by other users. Once a
post is created, it will exist on a user’s account indefinitely unless it is deleted. A story is
similar to a post, but it is only up for a short time and comments and likes are not
displayed. Instead, stories act as a slideshow that move through videos and photos posted
by the accounts a user follows. Others can comment on a story, but the comment and who
viewed a story is only visible to the creator. Currently, Instagram only supports mobile
users viewing the story of accounts that they do not follow.
8
Figure 1. An Instagram Post: 1) Media – can be a video or photo, 2) Owner Thumbnail,
3) Owner Username, 4) Post Caption, 5) Post Tags, 6) Commenter Username, 7)
Comment, 8) Post Likes, 9) Post Date
A user’s history of posts is stored as an account, shown in Figure 2. An account
page shows the user’s thumbnail and username. It also has statistics on how many posts
the account has made, how many other users follow the account, and how many others
the user follows. To follow an account means to have the posts of that account appear in
the stream of posts provided to the user, called their feed. If someone is signed into
Instagram, it also possible to view which accounts are followers and which accounts the
given account is following. Next, accounts have a bio that may describe their content or
themselves. Finally, on the account page are listed the posts of the user from most recent
to oldest.
9
Figure 2. An Instagram Account
Similar to an account page, Instagram has pages for locations. Locations are
created by users, and subsequently used by others to classify future posts. An example
location is shown in Figure 3. The page shows the coordinates of location, a story that
users can add content to, the current top nine posts, and the list of all other content in
reverse chronological order. Since locations are user created, there is not an extensive list
of all locations, but it is possible to view the top 1000 locations for a city by navigating
the options of the “explore/locations” subdomain.
10
Figure 3. An Instagram Location
2.4 Data Collection and Management
Data analysis can be thought of as a continuing process rather than a one-step
application. This analytic cycle is depicted in Figure 4. Although much of the work in the
process is done in the sub-section labeled “understand”, it is also important to not neglect
the first two steps, “import” and “tidy.” Importing entails pulling information from a
source, whether it is already organized or through more intensive processes like web
scraping. Tidying information is the process of formatting data into a usable and
consistent structure (Wickham & Grolemund, 2016). These two steps are essential to the
analytic process since without them, the analyst has no data to work with or the data
exists in a form that cannot be analyzed. Therefore, in the pursuit of analyzing social
media information, attention should be given to the importing and tidying of the data, as
is the emphasis of this thesis.
11
Figure 4. The Analytic Cycle (Wickham & Grolemund, 2016)
The first step to acquiring information from a social media site is determining an
avenue to the information. One source of information is to scrape the desired fields
directly from the displayed page, commonly referred to as web-scraping. Such a program
would traverse the actual webpage that is displayed when a person uses the site. To move
between pages, it would use built in links or travel along links in the page to search for
requested information. This web-scraping approach relies on the underlying structure of
HTML-based sites to process what was made for human interaction (Glez-Peña,
Lourenço, López-Fernández, Reboiro-Jato, & Fdez-Riverola, 2013). This approach offers
complications to a program since it does not adapt well to changes to the site and requires
specific implementation for each portion of a site (Glez-Peña et al., 2013). Most websites
also offer an application programming interface (API) which allows outside organizations
access to the information stored in the sites databases. APIs offer a more consistent
means of retrieving data in consistent formats, but it is not guaranteed that they will offer
all information that the site stores (Glez-Peña et al., 2013). Additionally, as is the case
with Instagram, APIs may be subject to strict regulations that prevent the public from
accessing them. The alternative to this is an open API, which allows third-party
12
developers access without needing authorization, but it can still suffer the other problems
of APIs (Jang & Lee, 2009). With both APIs and web scraping, sites can be behind an
authentication wall that will not allow some or all of their information to be displayed
unless the user is logged into an account.
Once a method of access to the website is determined, the information is still in
need of storage and management. The information collected is still in a raw state and as a
stream is difficult to work with (Injadat, Salo, & Nassif, 2016). When web scraping, the
data can be stored in almost any format, depending on how the web scraping software
pulls its information. APIs will return a data structure based on established guidelines,
one of the benefits of interacting through an API (Glez-Peña et al., 2013). In the case of
Instagram’s open API, data is return in JavaScript Object Notation (JSON). JSON files
are fundamentally text files, but they are organized in a way to make them both readable
for humans and easy for machines to parse (“JSON,” n.d.). A complete list of data
formats would be infeasible, but the basic principle is that data will come in a certain
format depending on the chosen importing method.
Whatever the method, data will be required to conform to certain requirements in
order to be compatible with the needs of tools and data storage later in the process. This
conversion of data falls under the category of “tidying” the data, or ensuring that the data
is in format that best facilitates future work(Wickham & Grolemund, 2016). Toward this
goal, the programming language R has resources that facilitate the “tidying” process. R is
an open-source programming language with a focus on statistical computing and
extensibility (“R: What is R?,” n.d.). R is extended via packages, fundamental units of
reproducible code (Hadley Wickham, 2015). One such package, tidyr, specializes in
13
tidying data (Hadley & Henry, 2018). This package and more like it make R a great
programming language for manipulating data of many different formats and preparing it
for use in analysis.
A question not yet answered is what information is important to pull? It is
impractical to have access to all the data flowing through social media sites without the
cooperation of the company behind them, especially for sites as large as Instagram. A
decision must be made then to choose what data to target and how to find the data
relating to that target by second order connections. While the ultimate use of information
and methods of analyzing are outside the scope of this thesis, exploring potential use
cases allows for the creation of a program that more accurately reflects the needs of
future users.
One strategy for analysis is a user focused approach. By viewing the post rate and
other statistics like following or resubmissions, analysts can put together a map of
influential users and their reach (Erlandsson, Bródka, Borg, & Johnson, 2016). Further,
they can track the flow of information and events to construct communities and networks,
similar to those created now with more conventional gathering techniques (Lv & Guo,
2016). Networks can also be established to connect users. Algorithms that watch the
behavior of users can flag behaviors and interests that share a high correlation with target
user bases, providing new insight into potential persons and groups of interest (Mezghani,
Péninou, Zayani, Amous, & Sèdes, 2017). This is an example of a unique application of
computers, where it may discover underlying connections that even an experienced
investigator might miss.
14
Another angle for analysis is to focus on events rather than networks. Social
media platforms have already shown the potential to be important barometers in public
perception of events and topics. By following trends, hashtags, and locations of posts,
analysists can create models that capture the sentiment and movement of a target
audience (Bian et al., 2016). This leads again to an area where computer driven analytics
can greatly help in information gathering. In a world of social media and
interconnectedness, discussion of events can evolve faster than a system administrator
can have time to adapt to. So it becomes important for the information gathering
software to have ways of dynamically associating common terms and pulling from
sources that a conventional search might overlook (Chan, Vasardani, & Winter, 2014).
2.5 Existing Instagram Data Extraction Software
Tools for extracting information from Instagram currently exist but are limited in
their usefulness. An overview of some available solutions is presented in Table 1. This
review will look more closely at instaR and Instagram PHP Scraper, but looking
at Table 1 shows that the various programs approach the same problem in many ways.
There are free programs like Instagram-scraper by Richard Arcega (Arcega, 2018)
and monetized software like the Instagram scraper provided by Im Rista (“Instagram
Scraper - Im Risto - Internet Marketing Blog,” n.d.). Many of the programs are based on
open source languages like Python and PHP, but only instaR notably uses R. The
monetized programs do not make their source code available, and as such, would not
easily be extendable. The existing software serves as a framework to inform
functionality, but ultimately each program fails to capture every important feature. The
monetized software has restrictive licensing issues and closed code. Programs in PHP and
15
Python do not capitalize on the existing analytic tools employed by the DoD community.
And lastly, instaR relies on interactions with Instagram’s private API, and the
authorization required to use that do not fit with mission needs.
Table 1 - Instagram Data Extraction Software: O – Optional, G – GitHub, ? –
Unavailable Information
insta
gram
-scra
per
Insta
gram
Super Scra
per
Insta
gram
PHP Scraper
Im R
isto
Insta
Scraper
insta
R
R ✓
Python ✓
PHP ✓ ✓
Custom ✓ ✓
API ✓ ? ✓
Scrapping ✓ ?
JSON ✓ ✓ ✓ ?
Authentication O O O ✓ ?
Authorization ✓
Availability G G G $49 $8 G
GUI ✓ ✓ ✓
These programs, however, can still be useful as templates for this research. First,
this thesis explores the instaR package by Pablo Barberá (Barberá, 2017). The program
is an R-based package that allows the user to interact with Instagram’s API to perform a
range of functions that are detailed in Table 2. The functionality of instaR is not as
expansive as other programs, but it highlights a few important considerations for
Instagram scrapers. Besides being outside of DoD control, instaR would have been a
good choice for Instagram data extraction based in R if it didn’t require authorization.
16
Since November of 2015, all API endpoints require a valid access_token, which requires
the application to be reviewed and approved by Instagram themselves (“Platform
Changelog • Instagram Developer Documentation,” n.d.). This is already a difficult
process for civilian users, but is a severe deterrent for the use of the API in a DoD
capacity. However, working with the API has two advantages Instagram-php-scraper, the
program investigated next, does not have. First, working with the API generally ensures a
higher level of stability with how a function will behave. Secondly, the API has some
functions that are hard to replicate through other means, namely, the ability to search for
posts in a given region.
Table 2 - instaR Functions
Function Purpose
getComments Retrieves up to 150 recent comments for a given post
getFollowers Retrieves the list of users that follow a given user
getFollows Retrieves the list of users a given user follows
getLikes Retrieves the list of users who liked a post
getLocation Retrieves location information
getPopular Retrieves up to 24 popular posts
getTagCount Retrieves a count of times a hashtag has been used
getUser Retrieves public information about a user
getUserMedia Retrieves public media from a given user and can
download media
instaOAuth Creates an OAuth access token enabling R to
communicate with Instagram’s API
searchInstagram Search media for mention of hashtag or for posts in a
given location
The other program investigated in detail is Instagram-php-scraper by
Postaddict.me (Postaddict.me, 2018), whose functions are detailed in Table 3. Unlike
17
Table 3 - Instagram-php-scraper Functions
Function Purpose
withCredentials Allows the program to run with account information
searchTagsByTagName Returns the hashtag results in a search for a given tag
getErrorBody An error handling function
searchAccountsByUsername Returns the accounts results in a search for a given
username
generateHeaders Adds credentials to a query
getMedias Retrieves the n most recent posts for a given username
getMediaById Retrieves information about a post with a given ID
getMediaByUrl Retrieves information about a post with a given URL
getMediaByCode Retrieves information about a post with a given
shortcode
getPaginateMedias Returns the first page of recent posts for a given
username
getMediaCommentsById Retrieves comments on a post with a given ID
getMediaCommentsByCode Retrieves comments on a post with a given shortcode
getMediaLikesByCode Retrieves likes on a post with a given shortcode
getAccountById Retrieves information on an account with a given ID
getAccount Retrieves information on an account with a given
username
getMediasByTag Retrieves n most recent posts with a given hashtag
getPaginateMediasByTag Returns the first page of recent posts with a given
hashtag
getCurrentTopMediasByTagName Retrieves the top nine posts with a given hashtag
getCurrentTopMediasByLocationId Retrieves the top nine posts for a given location ID
getMediasByLocationId Retrieves the n most recent posts for a given location ID
getLocationById Retrieves information for a given location ID
getFollowers Retrieves the accounts following a given account ID
getFollowing Retrieves the accounts a given account ID is following
getStories Retrieves the story information for a given account
setProxy Establishes a proxy for queries to be passed through
instaR, Instagram-php-scraper accesses Instagram’s open API. This allows users
to retrieve most public information without needing authorization from Instagram or even
an account. However, it does allow for users to log-in to access information limited to
authenticated accounts. The functionality of this program is also much more extensive
than others. Note, however, that while it has functions that interact with specific
locations, it cannot find posts in a given region. The two main factors that make
18
Instagram-php-scraper an imperfect candidate is the language it is coded in, PHP,
and that it is controlled by a source outside the DoD. It does, however, serve as an
excellent blueprint for other programs interacting with Instagram’s open API.
2.6 Reproducible Analysis
The last subject that this thesis aims to address is the reproducibility of analysis.
To the ends of validation and verification, is has been standard for independent bodies to
try and replicate the results of a study (R. D. Peng, 2011). However, with more research
being done with the aid of data and independently collected information, all sciences are
suffering a crisis of reproducibility (R. Peng, 2015). Pulling information from a website
can result in drastically different outputs if the procedures for searches and search criteria
are not specified. Even then, the information might have changed since most websites
allow edits and deletions, as well as new data always being entered. The concept of
ensuring that the data and procedures for a study are made available to others is referred
to as “reproducible research” (R. D. Peng, 2009). For the context of intelligence agencies,
this is valuable as an extra step of validation. One of the largest benefits of using open
source software is that it is highly reliable and reproducible (Ven & Verelst, 2006). When
future users attempt to reproduce the results of previous work, open source programming
languages like R allow them to ensure the same work is being done. In a practical sense,
this means that a user knows what to expect when they run a function. Unlike in closed
source software, a user could even look at the source code to verify that the same
operation is being carried out. In scholarly and scientifically rigorous applications,
developers should design software to facilitate reproducibility.
19
III. Methodology
3.1 Overview
The goals of this research are achieved by creating an R package called
instaExtract hosted on the GitHub page for the Data Science Lab headquartered in the
Air Force Institute of Technology. The package uses publicly available links from
Instagram to retrieve information on users, posts, and locations. Using R as the backbone
for this package allows analysts to easily handle the data returned as well as create
reproducible reports. Being hosted on the Data Science Lab’s GitHub page allows for the
package to be both readily accessible and routinely updated. As a demonstration of the
future of such a package this thesis includes an example function,
createLocationMapping(), to demonstrate what a future user might contribute to the
project.
3.2 A JSON-Based Instagram Scrapper in R– instaExtract
The limitations of other accessible Instagram scraping software have already been
discussed. In addition to being outside the military community’s reach of control, these
other programs suffer from challenges that make them unsuitable for use by intelligence
gatherers. Namely, they are limited by interacting with the restrictive and exclusive
Instagram API or they are difficult to operate and distribute.
This research overcomes these challenges with the creation of an R package that
is based on interactions with publicly available Instagram links that return JSON files.
The primary advantage of acquiring data this way is that most information on the site can
be obtained without the need to log-in, nor interact at all with the closed API. This would
allow a user behind a proxy to access almost all information on Instagram without
20
needing to provide any of their own information. Furthermore, the use of JSON links is
far less demanding on a user’s bandwidth than loading every page in a traditional
browser. A JSON link will not load images or videos which greatly reduces the time
demand of queries. Instead, all the information for a page is reduced to the data defining
it. For instance, an image would be represented by a unique identification number and
Uniform Resource Locator (URL). If, in the future, a user wanted to download an image,
they could use this information to download it.
3.2.1 Helper Functions
There is a collection of generic links that the package uses to create these JSON
links. To access the information a user wants, the package will take a user input and
substitute it into the generic link for each category. For more complex interactions, the
package will use information contained in the JSON file to create a new link, simulating
pagination through all the possible results, or until the desired number of results is
returned. These links and their manipulators are located in the Endpoints.R file. In the
example below, an example user query is used to create a working link to Instagram’s
open API.
User_Media_Json_Link <- "https://www.instagram.com/{username}/?__a=1&max_id={maxID}"
getUserMediaJsonLink <- function(user, maxID){ link <- User_Media_Json_Link link <- gsub("{username}", user, link, fixed = TRUE) link <- gsub("{maxID}", maxID, link, fixed = TRUE) return(link) }
21
To work with Instagram’s identification numbers, the package utilizes transformer
functions in Transformers.R. Instagram uses both IDs and shortcodes to identify their
content. IDs are most commonly returned in the JSON file, but shortcodes are the
identifiers used when creating the links. To go from shortcode to ID, for each character in
the code, the program multiplies a placeholder ID by 64 and adds that number to the
index of the given character in an alphanumeric alphabet. To go from ID to shortcode,
process is reversed. However, since Instagram’s IDs are larger than the integer size that
R works with, the package utilizes the bit64 package to accurately represent the integer
when doing math. The functions used to perform these operations are shown below.
getIDFromCode <- function(code){
alphabet <-
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_" ID <- 0
for(i in 1:nchar(code)){
c <- substr(code,i,i)
id <- id * 64+ regexpr(c, alphabet)[1]
}
return(ID) }
getCodeFromID <- function(ID){
alphabet <- "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_"
code <- ""
ID <- bit64::as.integer64.character(ID)
while(ID >0){
remainder <- ID %% bit64::as.integer64.double(64)
ID <- (ID - remainder) %/% 64 code <- paste(substr(alphabet,remainder+1,remainder+1),code,sep="")
}
return(code)
}
22
3.2.2 Search Functions
With endpoints and transformers established, the package can utilize a large
assortment of functions to pull specific information or search for a user provided query.
The smaller subset of functions are the search functions, which return tags or account
names similar to the query given by the user. Table 4 shows these two functions and
their purposes.
Table 4 - Search Functions Provided by the instaExtract Package
Function Purpose
searchAccountsByUsername Will return a data frame of accounts and their
information whose username contains the users
query
searchTagsByTag Will return a data frame of tags in use that contain
the users query
3.2.3 Get Functions
The majority of instaExtract’s functionality comes in the form of its Get
functions, which retrieve information based on what the user supplies. Table 5 lists these
functions as well as their purposes. All these functions operate without a user needing to
log in. When using a function, the user will typically provide two arguments, the query
they are interested in and the number of results they wish to return. In cases where the
number is not provided, the package will only return the first page of information. The
package will then return a data frame containing information relevant to the given
element in question. These results are detailed in Table 6.
23
Table 5 - Get Functions for the instaExtract Package
Function Purpose
getCommentsByMediaCode Return the first n comments and comment
information for a media with a given
shortcode
getCommentsByMediaID Calls getCommentsByMediaCode after
transforming a media ID into a shortcode
getCurrentTopMediaByLocationID Returns the top nine media for a location
with the given location ID
getCurrentTopMediaByTag Returns the top nine media for a given
hashtag
getLikesByMediaCode Returns information on the first n likes on a
media post with the given shortcode
getLikesByMediaID Returns information on the first n likes on a
media post with the given ID
getLocationByID Returns information on a location for a
given location ID
getMediaByCode Returns information on a media post for a
given media shortcode
getMediaByID Calls getMediaByCode after transforming
ID to a shortcode
getMediaByLocationID Returns information on the n most recent
media post for a given location ID
getMediaByTag Returns information on the n most recent
media post for a given hashtag
getMediaByURL Returns information on a media post with a
given URL
getMediaByUsername Returns the n most recent media posts from
an account with the given username
24
Table 6 - Format of the Results for the Get Functions in the instaExtract Package
Function Results
getCommentsByMediaCode n x 6 data frame: id, text, created_at, owner.id,
owner.profile_pic_URL, owner.username getCommentsByMediaID n x 6 data frame: id, text, created_at, owner.id,
owner.profile_pic_URL, owner.username getCurrentTopMediaBy LocationID
9 x 14 data frame: id, thumbnail_src, tubmnail_resources,
is_video, code, date, display_src, video_views, caption,
dimensions.height, dimensions.width, owner.id,
comments.count, likes.count
getCurrentTopMediaByTag 9 x 14 data frame: id, shortcode, taken_at_timestamp,
display_URL, tumbnail_src, tumbnail_resources, is_video,
video_view_count, edge_media_to_caption.edges,
edge_media_to_comment.count, dimensions.height,
dimensions.width, edge_liked_by.count, owner.id
getLikesByMediaCode n x 7 data frame: id, username, full_name,
profile_pic_URL, is_verified_followed_by_viewer,
requested_by_viewer
getLikesByMediaID n x 7 data frame: id, username, full_name,
profile_pic_URL, is_verified_followed_by_viewer,
requested_by_viewer getLocationByID 1 x 6 data frame: id, name, has_public_page, lat, lng, slug
getMediaByCode 1 x 42 data frame: meta information about a post not
including its comments and likes getMediaByID 1 x 42 data frame: meta information about a post not
including its comments and likes getMediaByLocationID n x 15 data frame: comments_disabled, id, thumbnail_src,
thumbnail_resources, is_video, code, date, display_src,
video_views, caption, dimensions.height,
dimensions.width, owner.id, comments.count, likes.count
getMediaByTag n x 16 data frame: comments_disabled, id, thumbnail_src,
thumnail_resources, is_vdeio, code, date, display_src,
caption, dimensions.height, dimensions.width, owner.id,
comments.count, likes.count getMediaByURL 1 x 42 data frame: meta information about a post not
including its comments and likes getMediaByUsername n x 17 data frame: _typename, id, comments_disabled,
getting_info, media_preview, thumnail_src,
thumbnail_resources, is_video, code, date, display_src,
caption, dimensions.height, dimensions.width, owner.id,
comments.count, likes.count
In order to provide a better understanding of how the package operates, this paper
provides a walk-through of the getMediaByTag function, which is representative of
25
how most of the other Get functions operate.
getMediaByTag <- function(tag, n = 20, maxID = ""){
The function call has three arguments: 1) tag – the query a user is looking to
gather data on; in this case, a hashtag. 2) n – the number of responses that will be
returned. Each function defaults to the number of results that are returned in the first
JSON file. 3) maxID – an identifier that allows the function to start its search from a
different position. The user will usually only use the default of an empty character, but
the function will use it to paginate through JSON results.
i <- 0
moreAvailable <- TRUE
data <- data.frame()
Next, the function initializes variables used in its operation. The integer i is used
to keep track of the number of results returned. The boolean moreAvailable will track
if more results from Instagram are available. Initializing data as a data frame allows
inserting new rows as results come in.
#will run while more data exists and it has not reached n results
while(moreAvailable && i < n){
The main operations of the function, creating a link and collecting the results, will
happen in this while loop. It will loop until there are no more results available or until
the number of results is greater than n.
26
#create the url from Json Link
url <- getTagMediaJsonLink(tag,maxID)
Tag_Media_Json_Link <- "https://www.instagram.com/explore/tags/{tag}/?__a=1&max_id={maxID}"
getTagMediaJsonLink <- function(tag, maxID){ link <- Tag_Media_Json_Link link <- gsub("{tag}", tag, link, fixed = TRUE) link <- gsub("{maxID}", maxID, link, fixed = TRUE) return(link) }
Calling the helper function getTagMediaJSONLink creates the JSON link URL.
When this function is called, it passes the desired hashtag and the current maxID. The
helper function will then alter a constant string that was defined before compilation by
inserting the given hashtag and maxID. The character string this function returns is the
URL that will be used to retrieve the desired JSON file.
#the unflattened response response <- jsonlite::fromJSON(url)
#will return as list if there is only one result
if(!is.data.frame(response$graphql$hashtag$edge_hashtag_to_media$edges$node)){ return(response$graphql$hashtag$edge_hashtag_to_media$edges$node) }
else{ #flattening the data down to the nodes, into a dataframe media <- jsonlite::flatten(response$graphql$hashtag$edge_hashtag_to_media$edges$node)
27
Next, the URL is used to retrieve the JSON file from Instagram. The imported
function jsonlite::fromJSON translates the JSON file to a R usable data frame. If
there is only one result in this JSON file, it will return this list without doing the rest of
the function which assumes that it has a data frame. In the normal case, when a data
frame is returned, jsonlite::flatten will flatten the data frame to be a regular two
dimensional data frame, meaning that each column of the data frame is atomic, rather
than a list with multiple values. In most cases, the desired information is located inside
containers such as “graphql” and coupled with superfluous information. To target the
useful information, the package must reference the correct container in the data frame
returned from the JSON link. In each case, this path is unique to a function and can
change between Instagram updates. This points towards a need to have this package
quickly modifiable and subsequently distributable, which is addressed in Section 3.3.
#iterating over the rows of the media for(row in 1:nrow(media)){
#will exit loop and return data if reaching the limit if(i == n){ return(data) }
#will add a new row of media to data data <- plyr::rbind.fill(data,media[row,])
#incrementing the counting index i <- i + 1 }
This segment of code will iterate through the rows of the results returned from
Instagram. Each time, it will add that row to the data frame intended to be returned. The
rbind.fill function from the plyr package to ensures that information from video
28
and picture posts can be added to the same data frame. The function keeps track of how
many rows it has entered and will return the results if it reaches the desired number.
#Where to start the next query to the instagram link
maxID <- response$graphql$hashtag$page_info$end_cursor
#makes sure more exists
moreAvailable <-
response$graphql$hashtag$page_info$has_next_page
}
}
The last step in the while loop is storing the identifier that allows the function to
retrieve the next set of information. Again, emphasis is given to the nature of the JSON
files’ structures returned. The location of these identifiers inside the JSON is subject to
change, and without a method for updating, the function might become obsolete.
#convert the json data to R dataframe return(data) }
Now that data is filled with n results, it is returned to the user. Each Get function
returns a variety of data frames with different elements, as outlined above in Table 6. The
remaining Get functions operate very similarly to getMediaByTag, however each has a
unique URL and data structure to its JSON file.
The Get functions in this package represent most of the functionality available in
other Instagram scrapping programs. Not currently possible is retrieving following and
follower information from an account, or getting a username from an owner ID. This gap
in functionality also means that the package operates without needing log-in credentials
or interaction with the Instagram API. Utilizing the JSON-based open API, the package
29
relies on constants that must be defined by the developer. Therefore, the maintainer of the
package is incentivized to house it in a manner that is adaptable to updates to Instagram.
This concern is solved by hosting the package on the AFIT Data Science Lab’s GitHub
page, a process discussed in the next section.
3.3 Adaptable, Reproducible, Distributable – Hosting on the AFIT Data Science
Lab’s GitHub Page
To ensure the instaExtract package is adaptable, reproducible, and
distributable, it is hosted on it on the GitHub page of the AFIT Data Science Lab. From
their website, “GitHub is a code hosting platform for version control and collaboration”
(“Hello World · GitHub Guides,” n.d.). That succinct description captures why GitHub
can be an indispensable tool for this package and for many military analysis operations.
GitHub allows the package to be accessible to all systems that have access to the internet,
to store previous versions of the package to reproduce reports done in the past, and to
facilitate discussion and evolution between users and developers of the package.
3.3.1 Benefits of GitHub for Military Analysts
An immediate consequence of hosting the package in an online repository is that
the most current edition of the package is available to any user with access to the internet.
The tools needed to conduct Instagram data scraping are not locked away on a machine.
Instead, the user can have a consistent and reliable way to locate the tools they are
familiar with and confident in. A further benefit of a centralized hub like the AFIT Data
Science Lab is that consolidated resources will lead to more capable and competent tools.
When the standard practice for analysts becomes working from a common set of
applications, those applications benefit from increased attention and use. Additionally,
30
users who begin working within established systems will have increased code and data
compatibility. The benefits of consolidation and cooperation are multiplicative.
GitHub furthers this research’s goal of enabling reproducible analysis. R as a
language already facilitates reproducibility with workable data types and dynamic reports
in R markdown. GitHub goes a step further and ensures that the exact version of the
package used to construct the report is available to an analyst who is validating it. GitHub
stores the changes made to a package, meaning as long as the version of the package used
is recorded in the report, it can be downloaded on the GitHub page. This prevents future
changes to a function from impacting the validation process. An older version of the
package should operate just as it did when the analysis was first conducted.
Perhaps the strongest benefit of using the GitHub platform is the level of
collaboration and adaptability. It is a common occurrence in large organizations, the
military included, for there to be a large disconnect between the developer of a tool and
the end users of that tool. The end user is often left with needless capabilities or lack
those that are essential. GitHub facilities a connection between users and developers, and
in some cases, blurs the distinction between the two. The development process for a
conventional program hosted on a local machine is filled with inefficiencies and
difficulties. First, the developer is made aware of a need. This alone might have taken
months to perforate a command structure and reach their desk. Next, they need to collect
information on the end user’s needs or operate off of the limited information they might
have been given with the assignment. After working on the project for a time, they will
seek feedback with the end user. This will lead to revision cycles that take time, money
and are bound to miss complications that will arise when the program ships. The resulting
31
program will then need a method of distribution, which can range from an online
download, to an executable passed through a file sharer. Whatever the method, the end
user is left with a program that might or might not meet all their needs. The first few
weeks of operations will be monitored by the developer, who can make necessary
changes and encourage users to download the most recent version. But after a time, the
gap between a program’s users and the developer widen. New issues arise, whether from
updated needs of the user or from unforeseen use cases. Complaints may be raised, but
resolutions suffer from the same slowdowns that the initial development suffered from. If
users result to fixing these issues themselves, different versions of the program can begin
to diverge, causing new issues with incompatibility and distribution. The end result of
the conventional set up is that developers are isolated from the users they are creating for
and the users are stuck with programs that are stagnant and restrictive.
Comparing that process with that of development through GitHub illustrates the
benefits of a shared hosting platform. With centralized hubs of development, like the
AFIT Data Science Lab, requests for new tools can be addressed with expedience.
Depending on the scope and the urgency of the problem, the development might be
conducted by students, civilians (government or non-government), or dedicated military
analysts. In all cases, the development process can be much more organic than the
conventional process. Errors and inadequacies are to be expected to some degree in all
forms of computer programs, but GitHub facilitates communication and issue resolution
to minimize the distribution flaws cause the user. In the conventional example, an issue
that arises after the development process of an application can be hard to remedy. With
GitHub, the end user can instantly notify the developer or the staff assigned as
32
administrators to the application. The error might be as simple as misuse of the
application, in which case, the administrator can reply to the issue and guide the user
through operation. If the issue turns out to be something more complex, the administrator
can use the issue as immediate feedback about the workings of the application. Since the
application is hosted in one predictable location, any updates the administrator makes to
fix the problem will automatically be reflected in the current of the program. Users can
always ensure that their programs are up to date and they are operating with the same
tools that everyone is using.
3.3.2 An Example of GitHub’s Potential Adaptability – Location Mapping in the
instaExtract Package
GitHub also offers users the ability to directly submit their own contributions,
blurring the line between developer and user and greatly increasing the adaptability of a
program. If a user finds that they have need for a new function, its within their power to
create it. GitHub gives user access to the source code, allowing them to add to and
modify their local version of the package. More than that, GitHub provides the means for
that user to add their own custom function to the main package. They may request a
change to the main program, and provided the administrators approve the change, that
custom function can now be used any other user with access to the package. This
integration of users into the development process results in ever expanding functionality
and increased resiliency.
To illustrate the potential of user created functionality, the instaExtract
package includes a set of functions that broaden the capability of the package and enable
users to conduct more thoughtful analysis. This set of functions is detailed in Table 7. A
33
current weakness in the JSON dependence of the package compared with the Instagram
API is that there is no readily available way to search for posts around a given latitude
and longitude. The solution presented by this research to this problem relies on extracting
the latitude and longitude from a location’s page. These values are then used to create a
location mapping, a data frame with locations and their information.
Table 7 - Functions Enabling Location Mapping Capabilities
Function Purpose
createLocationMapping Creates a data frame with all
locations in a search region
haversineDistance Calculates the distance in miles
between two latitude and longitude
points
getCurrentTopMediaByLocationMapping Returns the top n results for each
location in a location mapping
getMediaByLocationMapping Returns the n most recent results for
each location in a location mapping
getLocationsInRange Returns a filtered location mapping
with locations within a certain radius
The core function developed to achieve these functions is called
createLocationMapping.
createLocationMapping <- function(country ="", city = "", lat_long = FALSE){
This function creates the data frame in the format referred to as a “location
mapping” in this paper. It has three arguments: 1) country – the name/s of a country to
search inside. If no name is given, the mapping will be for all countries in Instagram’s
explore page. 2) city – the name/s of a city to search inside. If no name is given, the
mapping will be the top 1000 cities for all countries in the scope of the function. 3)
34
lat_long – if true, the function will gather the latitude and longitude values for each
location. It is defaulted to false since this operation requires a web query for each location
and can take significant amounts of time.
The key component to the operation of the createLocationMapping function
is the explore/locations page on Instagram. The top layer for this subdomain lists the
names of all countries in Instagram’s databases. The function will grab the names of all
countries in its scope, as well as their slug’s and ID. Using this information, it can move
down a layer to the country’s subpage, where it lists the top 1000 cities for that country.
Likewise, the function will grab information from this page to travel to each city within
the argument’s scope. The last level of explore/locations is a list of the top 1000 locations
in all the selected cities. These locations are what are added to the location mapping,
along with their slug and ID and the slug and ID of the city and country they belong to. If
the user selected to gather coordinates from the location, the function will also visit each
location page to grab the latitude and longitude. This can take a very long time depending
on the speed of the user’s internet connection and the number of locations in the location
mapping. However, since the information is retained in a data frame, a user can utilize
location mappings made in prior, avoiding this time constraint. Furthermore, it would be
possible for future contributors to build package that facilitate this process by hosting
current versions of location mappings. This would mean the analyst only has to
download the relatively small file instead of performing thousands of time consuming
queries.
With the location mapping created, it offers the user new manners for conducting
smart analysis on Instagram’s potential data. Two of the functions extend the package’s
35
ability to get media from a location to getting media from a location mapping. The
haversineDistance function enables the user to filter a location mapping to only
locations with a mileage radius of place of interest. To demonstrate what these functions
add to the package, this paper walks through an example scenario. Suppose that a user is
interested in what posts are being made in vicinity of the White House. The first step
would be to create the location mapping of Washington D.C.
washington_dc <- createLocationMapping("United States", "Washington", TRUE)
This function call creates the a mapping for the city “Washington” in the country
“United States” with the coordinates information. The variable washington_dc now
contains information for the top 1000 locations on Instagram in Washington DC.
lat <- 38.8977
long <- -77.0365
range <- .5 #miles
Next, the user sets the region they are interested in. The coordinates for the White
House are 38.8977° N, 77.0365° W. In this case, they are interested in locations within
half a mile from the point.
36
locations_near_white_house <- getLocationsInRange(washington_dc, range, lat, long)
getLocationsInRange <- function(mapping, r, lat, long, ...){
if(!is.numeric(r) || !is.numeric(lat) || !is.numeric(long)){ stop("r, lat, and long, must be numeric") } mapping <- filter(mapping, haversineDistance(latitude, longitude, lat, long) <= r)
return(mapping)
}
Using the simple function, getLocationInRange, the user can narrow down
the location mapping to a location mapping of only those locations within the range of
the white house set before. The haversineDistance function will return the distance
of each point from the White House, and only locations within half a mile from the White
House will be included in the location mapping named
locations_near_white_house.
near_house_media <- getMediaByLocationMapping(locations_near_white_house)
The last step is to use location mapping version of the get media function to
retrieve information on the recent media from the locations within our range. This
procedure results in a data frame with information on the most recent posts made at
locations near the White House. A user would be free to conduct whatever analysis they
wish with the data provided, but possible areas of interest might be mapping commonly
used hashtags for an area, comparing data from multiple days to observe fluctuations in
time, identifying prolific accounts, etc.
37
The location mapping functions created represent the future potential of the
instaExtract package or any package hosted on a collaborative platform like GitHub.
As users discover new ways to conduct analysis or require new data sources, the
functions they spur or create themselves can be added to the package. This in turn,
facilitates and inspires other analysts in conducting sophisticated and modern analysis of
their own.
3.4 Summary
This research culminates in a R package called instaExtract hosted on the
AFIT Data Science Lab’s GitHub page. R is an open source program that is easily
accessible and built around reproducible practices. The instaExtract package
provides users with a way to scrape current Instagram data without authentication or
interaction with the Instagram API. By hosting the package on the AFIT Data Science
Lab’s GitHub page, the package has an increased potential to stay relevant and useful to
real world analysts. Also included is an original set of functions curtained around
mapping locations to illustrate how future collaboration on this project might continue to
extend its usefulness.
38
IV. Exemplar Analysis
4.1 Overview
To better illustrate the capabilities and value of the instaExtract package, this
thesis includes an exemplar analysis that will emulate how an end user might use the
package. The user is assumed to have a working understanding of R and Instagram. The
first scenario focuses of an analyst investigating threats near a location, in this example
the White House. Section 4.2 discusses the gathering of information about the area. Next,
the analysis will conduct basic filtering and visualization techniques as well as other
instaExtract functions to investigate further.
4.2 Gathering Information on the Region
The investigation of the threat begins by collecting information about the
locations surrounding the White House. The user creates a location mapping of the
Washington D.C. region. The first 30 entries of the data frame of 1000 locations
generated by this function are found in Appendix B.
washington_dc <- createLocationMapping("United States", "Washington", TRUE)
Next, the location mapping is narrowed down to find only the locations within a
mile of the White House. As in the demonstration in the methodology section, the user
sets the coordinates of the White House and use the function getLocationsinRange
to filter the locations. The first 30 entries of the data frame of 524 locations are shown in
Appendix C.
39
lat <- 38.8977
long <- -77.0365
range <- 1 #miles
locations_near_white_house <- getLocationsInRange(washington_dc, range, lat, long)
Again, the next step is to collect the media from the area, but to be a little more
thorough, this user will get the last 50 posts from each area. The first 60 of 26,200 results
are partially shown in Appendix D. The complete values of the URLs and captions are
obscured to allow the table to be represented on one page. This operation took about 30
minutes, meaning that instaExtract is able to pull the information about a post around
15 times a second.
media_near_white_house <- getMediaByLocationMapping( locations_near_white_house, 50)
4.3 Hashtag Investigation
Now possessing media data, the user can begin to investigate it. There are many
different avenues to take at this point, where the following courses represent only some
of the possible paths. First, the user chooses to investigate the hashtags being used. The
hashtags contained in the captions are isolated and sorted to find the most commonly
used hashtags. Using the tidytext and stringr packages, the user first breaks down
each caption to separate elements for each word. Then, the words are filtered down to
only those words that start with ‘#’; our hashtags. Sorting these, the user finds the most
common hashtags recently used in the search region. The results indicate that there are
36971 hashtags used in total, with the top 15 displayed below. Filtering that result
further, only 59 hashtags have been used more than 100 times in search radius. Further,
40
as Figure 5 and Figure 6 show, even the top 150 hashtags are greatly skewed to the top,
with the top two, “#washingtondc” and “#dc” being used upwards of 2000 times each,
with next highest, “#washington” only being used 513 times.
library(stringr)
library(tidytext)
hashtags <- media_near_house3 %>% unnest_tokens(word, caption,token = "regex") %>% filter(substr(word,1,1) == '#') %>% count(word,sort = TRUE) # A tibble: 15 x 2 word n <chr> <int> 1 #washingtondc 2612 2 #dc 2171 3 #washington 513 4 #travel 489 5 #usa 430 6 #igdc 423 7 #love 405 8 #instagood 371 9 #acreativedc 369 10 #tbt 350 11 #photography 338 12 #foodie 325 13 #foodporn 313 14 #washingtonmonument 296 15 #valentinesday 295
42
Figure 6. Count of 150 Most Used Hashtags in the DC Area
For the next data utilization scenario, the user has a list of key words that they are
monitoring. It is possible to compare the list of keywords against the list of hashtags to
look for suspicious hashtags. The code to perform this comparison and the top 6 results of
such a search are shown below. It reveals 148 potentially threating hashtags. The top
result, “#justgoshoot”, has been used 45 times in the scope of the search.
43
keywords <- c('shoot', 'gun', 'kill', 'bomb')
threats <-hashtags %>% filter(str_detect(word, paste(keywords, collapse = "|"))) # A tibble: 6 x 2 word n <chr> <int> 1 #justgoshoot 45 2 #photoshoot 34 3 #shoot2kill 10 4 #citykillerz 7 5 #shootingwithshooters 7 6 #tonekillers 7
This hashtag turns out to be a hashtag used by photographers, but if it really was a
threat, instaExtractcan continue provide more information. To find more data on the
usage of this hashtag, the user retrieves the last 10,000 posts that used this hashtag.
Further, to know who used this hashtag most in the most recent 10,000 uses, the user
counts the list again, and finds that the account with ID 6961358493 used the hashtag 50
times. Outside tools could provide for even more complex analysis. The large amount of
written word in the captions makes a good source for sentiment analysis. This data
contains numbers of likes and comments, so a user could also sort for popular posts that
fit a given criteria.
44
just_go_shoot <- getMediaByTag("justgoshoot", 10000)
just_go_shoot %>% count(owner.id, sort =TRUE)
owner.id n <chr> <int> 1 6961358493 50 2 2014353892 45 3 495504345 37 4 7129727912 30 5 29197162 29 6 4843411210 29 7 374142894 28 8 298064835 24 9 1524025234 20 10 4976873299 18
4.4 User Investigation
A user can pull information from an account to find more detailed information
about that account’s history. The first step would be to download media from the chosen
account’s page. In this case, the account under investigation has the username
“usairforce.” Retrieving their last 2000 posts returns 1390 values, the first 60 of which
are shown in Appendix E, since this account has only posted 1390 times.
usairforce <- getMediaByUsername("usairforce", 2000)
The data collected can be used to get a better idea user’s is attributes. The package user
begins by finding the post with the most likes, their most popular post. Running any code
is not needed to do this, as R’s data viewer allows sorting a column in descending value.
In this case, the account’s top post, ID 1698060742908836418, has 41,541 likes, and the
photo is shown in Figure 7. The number of likes plotted against the postdate results in
45
Figure 8, showing the trend of like counts over time. This figure which shows a gradual
increase in the average likes over time.
Figure 7. Most liked Instagram Post by usairforce (usairforce, 2018)
46
Figure 8. Post Likes Over Time for usairforce Account
The package also allows a closer look at a post. A media ID is used to retrieve the
comments from the top post on this account in the getCommentsByMediaID function,
the first 60 of 269, which are shown in Appendix F. Using this information, the most
common words used in the captions are found in a process similar to that used above.
47
comments <- getCommentsByMediaID('1698060742908836418', n = 300)
commentWords <- comments %>% unnest_tokens(word, text,token = "regex") %>% filter(!word %in% stop_words$word) %>% count(word,sort = TRUE)%>% mutate(word = reorder(word, n)) # A tibble: 6 x 2 word n <fctr> <int> 1 love 14 2 air 13 3 rdgjklcb 11 4 force 10 5 girl 7 6 pilot 7
Comments can also provide insight to connections between users. Using the
comments from posts, the user can build a network of users who interact with the
usairforce account. To begin, 300 comments from the last 50 posts are collected. From
these, the users who have left the most comments is calculated. Although simple, the
network built from this data, Figure 9, hints at how networks can be developed from the
social media data obtained this way.
48
all_comments <- data.frame()
for(row in 1:50){
comment_holder <- getCommentsByMediaID(usairforce[row,'id'],300)
all_comments <- plyr::rbind.fill(all_comments, comment_holder)
}
commenting_users <- all_comments %>% count(owner.username, sort=TRUE)
# A tibble: 6 x 2
owner.username n
<chr> <int>
1 my_babel_physics_project_ai 30
2 admininnotinuse 28
3 zafarwestern 26
4 gomezbaquerosol 21
5 viktoria__lg.troxell 21
6 lydiavassallo 20
Figure 9. Network of Commenting Users on usairforce Posts Where Edge is
Proportional to Number of Comments
49
4.5 Geo Mapping
Part of the analytic cycle, shown in Figure 4, is visualizing the data. This allows
the analyst to grasp what type of data they are working with and understand the
significance of their analysis. R has many great packages that can help an analyst
accomplish this feat. One of these packages is leaflet, a JavaScript library used to
make interactive maps (Cheng, 2017). This section revisits the data used in section 3.3.2,
the location mapping of Washington DC. Using leaflet, an analyst can overlay the
locations in the location mappings with a real map of the area. The result, seen in Figure
10, is cluttered beyond much helpfulness.
Figure 10. All Washington DC Locations from Instagram Overlaid with Actual
Location
50
Luckily, leaflet comes with a clustering option that drastically increase the
readability of the map. In Figure 11, clustering has been turned on. The analyst can now
see where the locations are focused, without overwhelming the senses. Hovering over a
cluster shows the region it represents. For closer detail, the analyst can click on a cluster
to zoom in and expand its contents. Once a node is clicked, leaflet will display the name
of the location. Figure 12 is a fully zoomed in example. All of the data displayed in these
maps is nothing more than the information in a location mapping, but displaying that
information in a real world setting adds context and meaning to the information that
allows analysts to draw connections and make inferences that they would be hard pressed
to do without that context.
Figure 11. Clustered Locations for Washington DC
51
Figure 12. Zoomed in View of White House Instagram Locations
4.6 Summary
The applications of the data derived from this package are plentiful. With the
many R packages at the disposal of users, visualization and analysis can be accomplished
with style and ease. The data provided allows for media, user, and location driven
analysis. As this package matures, these procedures can be added to the functionality of
the package, but it currently serves an important role of providing useful and unique
social media data to analysts in the intelligence community.
52
V. Conclusions and Recommendations
5.1 Overview
The potential for the instaExtract package and the practices this thesis hopes
to encourage in the military analytical community has been shown to be promising, but
the long-term success of this and future programs will rely on the adoption rate of users
and the continued support of technical organization like the AFIT Data Science Lab. At
its core, the instaExtract package achieve its operational goal of providing resources
for analysts to extract Instagram data. Additionally, this thesis attempts to accomplish this
in a manner with forward reaching implications. R is a heavily supported, open source
language, with new resources created by the community on a routine basis. Choosing to
host the package on GitHub allows the package the benefit of future support and user
involvement.
5.2 The instaExtract Package
The instaExtract package boasts most of the same functionality offered by
similar available programs while exhibiting even more paths for it to grow. First and
foremost, the package achieves its goal of scraping Instagram for data. It offers search
functions and a suite of get functions that allow a user to obtain almost any information
that Instagram tracks. Furthermore, it has a collection of location mapping functions that
mimic the features that were traditionally limited to programs interacting with the
Instagram API. All of these functions can be operated without authorization and the need
to have log-in information for an Instagram account.
The package also has many areas for future development. While most information
can be retrieved without authorization, or otherwise inferred from available data, there
53
are some areas of Instagram that require log-in credentials. Future work should be done
to ensure the robustness of operations, such as compilation and use time error and
compatibility checking. As a final step towards professionalization, the package could
also be published to the Comprehensive R Archive Network (CRAN). CRAN is, as its
name suggests, a comprehensive archive of R packages that amount to the most
commonly used and relied upon functions in R. Publishing a package to CRAN allows
users to be sure of a certain set of requirements and accessibility that are enforced by the
CRAN maintainers.
In addition to work that falls within the purview of good coding practices, the
instaExtract package has many avenues of growth that would benefit users and
increase its applications. While certainly not exhaustive, this thesis suggests a few ideas
considered to be worthy of attention:
• Extending Location Mapping Functionality – Currently, the location mapping
functions can only access those locations that are listed through the
explore/locations page on Instagram. Adding functions that allow users to
quickly add their choice of locations to a mapping or that seek out other locations
on their own could further enhance the usefulness of this collection.
• A Shiny Application – Shiny is an R package that creates applications to allow
users to interact with a user interface rather than command lines in R. A Shiny
app would increase the reach of the package and potentially help inform its users
of all of its functions
• Proxy Settings – The instaExtract package uses the cURL package to
interface with internet. As such, it does not have well behaved proxy settings.
54
Incorporating better ways to control these settings could allow for a user to work
through a proxy, integrating with a local network or masking the location of their
requests.
• Other Scrapping Techniques –How the JSON link approach to data scrapping is
susceptible to changes in Instagram’s structuring has already been discussed.
However, it would be possible for future developers to allow for users to scrape
their selected data through other means. One such option is the Selenium
package, which uses conventional browsers to navigate between webpages.
Although it is exclusive, gaining access to the Instagram API could enable many
new options and considerably accelerate existing ones.
• Analysis Functions – The package was designed around the goal of scrapping
data, but it would be natural to include analytical functions in the future. The
current package lends itself to sentiment analysis, anomaly detection, and more.
With methods of pulling the images that are represented currently only as URLs
would also allow for image recognition.
5.3 The AFIT Data Science Lab
A major academic pursuit of this paper is to validate the viability and incentives
of programs hosted on a centralized and collaborative platform, such as the AFIT Data
Science Lab. The GitHub platform will greatly simplify and facilitate the future
development of this package. Users not only have access to a reliable and up-to-date
version of the software, but can maintain a responsive dialog with the maintainers of the
code as well as share the results of their individual efforts. GitHub’s version control also
55
enables analysis with a focus on reproducibility by storing previous versions of the
package.
The long-term value of these pursuits will be determined by the adoption rate of
similar practices by other analytical software and the amount of support given to software
development organizations. As discussed, packages stored in a centralized location
receive a multiplicative benefit of enjoying a higher user base and more engaged
developers. But of course, this requires the support structure to address development
issues and to create new packages to meet user needs. Ensuring this support staff is
valued and funded will require further commitment to similarly focused programs and
research that can validate the worth of such programs.
5.4 Summary
It is clear that instaExtract package adds value to the analytical and
intelligence communities, but this research acknowledges that there is more work to be
done in solidifying the robustness of this package and validating and realizing a
collaborative and comprehensive hub for analytic resources. Future research should be
done into more data acquisition software and the statistical benefits of platforms like
GitHub and their effects on DoD workflows.
56
Appendix A
Link to the instaExtract package: https://github.com/AFIT-R/instaExtract
57
Appendix B
id name slug city_ID city_Name city_Slug
country
_ID
country_N
ame country_Slug latitude longitude
213480180
Washington, District of
Columbia
washington-district-
of-columbia c2427178 Washington
washington-
united-states US
United
States united-states 38.8951 -77.0367
225931565 The Obama White House
the-obama-white-
house c2427178 Washington
washington-
united-states US
United
States united-states 38.89768 -77.03655
15712 Lincoln Memorial lincoln-memorial c2427178 Washington
washington-
united-states US
United
States united-states 38.889444 -77.050278
3001994 United States Capitol united-states-capitol c2427178 Washington
washington-
united-states US
United
States united-states 38.8897301 -77.0070362
214773851
Washington Monument
National Monument
washington-
monument-national-
monument c2427178 Washington
washington-
united-states US
United
States united-states 38.8890235 -77.0331092
4366681 National Gallery of Art
national-gallery-of-
art c2427178 Washington
washington-
united-states US
United
States united-states 38.8913397 -77.0196344
82474402
Smithsonian’s National Zoo
and Conservation Biology
Institute
smithsonians-
national-zoo-and-
conservation-biology-
institute c2427178 Washington
washington-
united-states US
United
States united-states 38.9299778 -77.0511297
1191441824276880 The White House the-white-house c2427178 Washington
washington-
united-states US
United
States united-states 38.8968447 -77.0366049
235453813 Nationals Park nationals-park c2427178 Washington
washington-
united-states US
United
States united-states 38.8732565 -77.0075808
372247132 Capital One Arena capital-one-arena c2427178 Washington
washington-
united-states US
United
States united-states 38.89795 -77.02096
262515071
Smithsonian National
Museum of African American
History and Culture
smithsonian-
national-museum-of-
african-american-
history-and-culture c2427178 Washington
washington-
united-states US
United
States united-states 38.8911015 -77.0325639
373555 Renwick Gallery renwick-gallery c2427178 Washington
washington-
united-states US
United
States united-states 38.898867 -77.039447
236471522 The Mall (Washington DC)
the-mall-washington-
dc c2427178 Washington
washington-
united-states US
United
States united-states 38.89 -77.0236111
214720506 The Capitol, Washington D.C.
the-capitol-
washington-dc c2427178 Washington
washington-
united-states US
United
States united-states 38.8898395 -77.0154594
279173 Howard University howard-university c2427178 Washington
washington-
united-states US
United
States united-states 38.92314 -77.02158
214638086
Smithsonians National
Museum of Natural History
smithsonians-
national-museum-of-
natural-history c2427178 Washington
washington-
united-states US
United
States united-states 38.8910781 -77.026232
2222215 National Building Museum
national-building-
museum c2427178 Washington
washington-
united-states US
United
States united-states 38.89779 -77.01752
838999
National Museum of American
History
national-museum-of-
american-history c2427178 Washington
washington-
united-states US
United
States united-states 38.8913741 -77.0299286
175770
Hirshhorn Museum and
Sculpture Garden
hirshhorn-museum-
and-sculpture-
garden c2427178 Washington
washington-
united-states US
United
States united-states 38.8883276 -77.0229156
212896512
National Air and Space
Museum, Smithsonian
Institution
national-air-and-
space-museum-
smithsonian-
institution c2427178 Washington
washington-
united-states US
United
States united-states 38.8881412 -77.0198422
42620 9:30 Club 930-club c2427178 Washington
washington-
united-states US
United
States united-states 38.91803 -77.02363
214513963
Union Station, Washington
D.C.
union-station-
washington-dc c2427178 Washington
washington-
united-states US
United
States united-states 38.8975 -77.00621
2150339
The George Washington
University
the-george-
washington-
university c2427178 Washington
washington-
united-states US
United
States united-states 38.8982042 -77.050286
139456 Jefferson Memorial jefferson-memorial c2427178 Washington
washington-
united-states US
United
States united-states 38.881111 -77.036667
576233589
Lijst van bekende mensen uit
Washington D.C.
lijst-van-bekende-
mensen-uit-
washington-dc c2427178 Washington
washington-
united-states US
United
States united-states 38.8951111 -77.0366667
216550363
National Portrait Gallery,
Smithsonian Institution
national-portrait-
gallery-smithsonian-
institution c2427178 Washington
washington-
united-states US
United
States united-states 38.8977641 -77.0229628
849479 The Library of Congress
the-library-of-
congress c2427178 Washington
washington-
united-states US
United
States united-states 38.8887863 -77.0058375
1009997177 Georgetown, DC georgetown-dc c2427178 Washington
washington-
united-states US
United
States united-states 38.9026909 -77.0494537
251483 Capitol Hill capitol-hill c2427178 Washington
washington-
united-states US
United
States united-states 38.889722 -77.011111
58
Appendix C
X id name slug cit y_ID cit y_Name cit y_Slug
count r
y_ID
count ry
_Name count ry_Slug lat it ude longit ude
1 213480180
Washingt on, Dist r ict of
Columbia washingt on-dist r ict -of -columbia c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.8951 -77.0367
2 225931565 The Obama Whit e House t he-obama-whit e-house c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89768 -77.03655
3 15712 Lincoln Memorial lincoln-memorial c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.889444 -77.050278
5 214773851
Washingt on Monument
Nat ional Monument
washingt on-monument -nat ional-
monument c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.8890235 -77.0331092
8 1191441824276880 The Whit e House t he-whit e-house c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89684467 -77.03660488
10 372247132 Capit al One Arena capit al-one-arena c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89795 -77.02096
11 262515071
Smit hsonian Nat ional
Museum of Af r ican
American Hist ory and
Cult ure
smit hsonian-nat ional-museum-
of -af r ican-american-hist ory-
and-cult ure c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89110149 -77.03256389
12 373555 Renwick Gallery renwick-gallery c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.898867 -77.039447
13 236471522
The Mall (Washingt on
DC) t he-mall-washingt on-dc c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89 -77.02361111
16 214638086
Smit hsonians Nat ional
Museum of Nat ural
Hist ory
smit hsonians-nat ional-museum-
of -nat ural-hist ory c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89107814 -77.02623198
18 838999
Nat ional Museum of
American Hist ory
nat ional-museum-of -american-
hist ory c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89137413 -77.02992863
19 175770
Hirshhorn Museum and
Sculpt ure Garden
hirshhorn-museum-and-
sculpt ure-garden c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.88832764 -77.02291558
23 2150339
The George Washingt on
Universit y
t he-george-washingt on-
universit y c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.8982042 -77.05028604
25 576233589
Lijst van bekende
mensen uit Washingt on
D.C.
lijst -van-bekende-mensen-uit -
washingt on-dc c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89511111 -77.03666667
26 216550363
Nat ional Port rait
Gallery, Smit hsonian
Inst it ut ion
nat ional-port rait -gallery-
smit hsonian-inst it ut ion c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.8977641 -77.02296276
28 1009997177 Georget own, DC georget own-dc c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.9026909 -77.0494537
30 2091823
Nat ional World War II
Memorial nat ional-world-war- ii-memorial c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.88935278 -77.04055556
31 26131 Newseum newseum c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89313 -77.01935
33 590718213 Dupont Circle dupont -circle c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.9096 -77.0434
34 21614
Walt er E. Washingt on
Convent ion Cent er
walt er-e-washingt on-
convent ion-cent er c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.9042406 -77.02325315
38 342980053 Tidal Basin t idal-basin c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.8837 -77.0389
39 215073 Ref lect ing Pool ref lect ing-pool c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.88926578 -77.04750981
40 683422 Madame Tussauds DC madame-t ussauds-dc c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89759 -77.0261699
41 75367954 Le Diplomat e le-diplomat e c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.91136386 -77.03157239
43 21783
Int ernat ional Spy
Museum int ernat ional-spy-museum c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89699 -77.02329
44 218723464
Nat ional Museum of
Women in t he Art s
nat ional-museum-of -women-in-
t he-art s c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.90006 -77.02916
47 212897059 Washingt on, DC washingt on-dc c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.89511111 -77.03666667
53 1420252
Mart in Lut her King, Jr.
Nat ional Memorial
mart in- lut her-king-jr-nat ional-
memorial c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.88611111 -77.045
55 11195531
Smit hsonians Freer and
Sackler Galler ies
smit hsonians-f reer-and-sackler-
galler ies c2427178 Washingt on
washingt on-
unit ed-st at es US
Unit ed
St at es unit ed-st at es 38.88796 -77.02645
59
Appendix D
com
me
nts
_dis
able
did
thu
mb
nai
l_sr
cis
_vid
eo
cod
ed
ate
dis
pla
y_sr
cca
pti
on
vid
eo
_vie
ws
dim
en
sio
ns.
he
igh
td
ime
nsi
on
s.w
idth
ed
ge_m
ed
ia_p
revi
ew
_lik
e.c
ou
nt
ow
ne
r.id
com
me
nts
.co
un
tli
kes.
cou
ntlo
cati
on
_nam
elo
cati
on
_id
1FA
LSE
1715
7344
3098
9320
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
b63
c603
9d2d
1715
eaa
dc9
f274
ee
3f45
/5B
167B
10/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
0.89
.108
0.10
80/2
7877
936_
1554
6539
5155
696_
7885
9276
2613
1374
080_
n.j
pg
FALS
EB
fPg1
_oB
o3N15
1875
1504
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
072c
0d0d
c9e
b07
3de
c7b
95f0
50f6
345/
5B24
7FD
E/t5
1.28
85-1
5/e
35/2
7877
936_
1554
6539
5155
696_
7885
9276
2613
1374
080_
n.j
pg
This
is R
adle
y. H
e is
a v
ery
go
od
bo
y. #
do
gso
fin
stag
ram
NA
1258
1080
020
9427
535
00
Was
hin
gto
n, D
istr
ict
of
Co
lum
bia
2134
8018
0
2FA
LSE
1715
7344
2283
6550
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
9e9d
7aca
e24
e37
a7e
eaa
59fb
2b17
a5e
/5B
1893
92/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/2
7581
389_
1869
0055
5410
992_
5265
6084
3005
9520
0_n
.jp
gFA
LSE
BfP
g14C
FTfT1
5187
5150
3h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/408
077b
6d4f
b3d
2e2b
972d
941a
7409
a2/5
B14
F8A
C/t
51.2
885-
15/e
35/2
7581
389_
1869
0055
5410
992_
5265
6084
3005
9520
0_n
.jp
gA
hh
hh
h -
on
e o
ur
very
co
zy T
ile
Bar
bin
de
rs g
ot
to s
nu
ggle
wit
h t
he
@ve
rve
coll
abo
rati
ve t
eam
mas
cot
Bah
hh
h-b
ara
#pap
araz
zi #
tile
bar
#d
cde
sign
#n
ew
add
itio
n #
luck
ybin
de
r #g
reat
gro
up
#ti
leb
ard
mv
#til
eN
A10
8010
800
5599
7855
360
0Th
e O
bam
a W
hit
e H
ou
se22
5931
565
3FA
LSE
1715
7344
2197
3230
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
41e
9f2b
6da9
1076
f509
7b63
7306
c915
/5B
2268
4D/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
0.13
4.10
80.1
080/
2757
7042
_348
8759
7893
2993
_529
3757
7616
8162
0992
_n.j
pg
FALS
EB
fPg1
3On
-La15
1875
1503
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
42c3
c1b
eff
4b9f
997b
4a74
390f
a690
6/5B
1FFD
EF/t
51.2
885-
15/e
35/2
7577
042_
3488
7597
8932
993_
5293
7577
6168
1620
992_
n.j
pg
Ugl
y cu
te ic
e c
ream
--N
A13
4910
800
5437
3333
90
0Li
nco
ln M
em
ori
al15
712
4FA
LSE
1715
7343
2827
1770
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
81a1
83b
0dcc
444f
dd
cf9a
b9a
c742
350/
5B23
3EFD
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c0.
86.1
080.
1080
/278
7883
6_17
2716
3927
3439
24_5
0236
6604
1738
2318
08_n
.jp
gFA
LSE
BfP
g0f9
lSo
m1518
7514
92h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/4fb
3719
a617
52f4
d8f
4217
90c5
8683
ca/5
B00
C93
6/t5
1.28
85-1
5/e
35/2
7878
836_
1727
1639
2734
3924
_502
3666
0417
3823
1808
_n.j
pg
NA
NA
1253
1080
021
7696
280
0W
ash
ingt
on
Mo
nu
me
nt
Nat
ion
al M
on
um
en
t21
4773
851
5FA
LSE
1715
7339
1507
4260
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/f
d1c
6cac
80d
0c61
56a9
e0a
3f3b
7fe
f65/
5B11
A68
7/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/2
7879
178_
1625
9684
7866
709_
6055
1809
8852
9508
352_
n.j
pg
FALS
EB
fPgu
fJF2
xU151
8751
442
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
2040
fd2e
87c9
adc7
b65
d11
1236
7470
c/5B
033D
E6/t
51.2
885-
15/e
35/2
7879
178_
1625
9684
7866
709_
6055
1809
8852
9508
352_
n.j
pg
Nat
ura
l lig
ht
and
fil
m g
rain
N
A10
8010
8011
1170
1555
350
11Th
e W
hit
e H
ou
se11
9144
1824
2768
80
6FA
LSE
1715
7338
4909
6440
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
30d
8d30
8a82
aeb
ab20
de
7bd
3951
d11
3/5B
1DA
44F/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c0.
135.
1080
.108
0/27
5772
45_3
2350
6974
8239
65_6
9461
7118
5285
5623
68_n
.jp
gFA
LSE
BfP
gth
sgc3
f1518
7514
34h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/8ac
ac2b
9c00
c566
b8e
e6c
11e
902b
c620
/5B
1090
9D/t
51.2
885-
15/e
35/2
7577
245_
3235
0697
4823
965_
6946
1711
8528
5562
368_
n.j
pg
He
re w
e a
re a
gain
. N
A13
5010
801
4689
1768
970
1C
apit
al O
ne
Are
na
3722
4713
2
7FA
LSE
1715
7337
7528
6350
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/f
5919
c016
7045
da2
07d
c610
1e7f
554a
6/5B
0A7C
C7/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/278
9205
8_41
5499
4655
4105
7_66
4604
0721
7719
8694
4_n
.jp
gFA
LSE
BfP
gsc9
FTV
e1518
7514
26h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/963
2fc0
c8b
19b
6f5a
c997
9e8c
7020
d69
/5B
044A
A6/
t51.
2885
-15/
e35
/278
9205
8_41
5499
4655
4105
7_66
4604
0721
7719
8694
4_n
.jp
gst
uff
y ca
ity
fro
m @
the
_sh
elt
er_
pu
ps
than
ks @
ahka
rp!
#ad
op
tdo
nts
ho
pN
A10
8010
802
1917
442
02
Smit
hso
nia
n N
atio
nal
Mu
seu
m o
f A
fric
an A
me
rica
n H
isto
ry a
nd
Cu
ltu
re26
2515
071
8FA
LSE
1715
7336
8261
7030
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
a317
e6b
299b
8978
30ca
8875
4198
aec0
/5B
0A95
64/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
135.
0.81
0.81
0/27
5750
78_1
5234
1312
0955
85_8
5301
6654
8663
3697
28_n
.jp
gFA
LSE
BfP
grG
pj6
NE15
1875
1415
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
9792
ce50
48d
b1c
ae09
8a6d
860d
e93
8b/5
B0E
166E
/t51
.288
5-15
/e35
/275
7507
8_15
2341
3120
9558
5_85
3016
6548
6633
6972
8_n
.jp
g<U
+041
0> <
U+0
441>
<U+0
435>
<U+0
433>
<U+0
43E>
<U+0
434>
<U+0
43D
><U
+044
F> <
U+0
440>
<U+0
435>
<U+0
448>
<U+0
438>
<U+0
43B
> <U
+043
F><U
+044
0><U
+043
8><U
+043
3><U
+043
E><U
+044
2><U
+043
E><U
+043
2><U
+043
8><U
+044
2><U
+044
C>
<U+0
428>
<U+0
430>
<U+0
443>
<U+0
440>
<U+0
43C
><U
+044
3>!!
!<f0
><U
+009
F><U
+009
2><U
+00A
A><
f0><
U+0
09F>
<U+0
08F>
<U+0
0BD
><f0
><U
+009
F><U
+008
F><U
+008
B><
f0><
U+0
09F>
<U+0
08F>
<U+0
0BC
><f0
><U
+009
F><U
+009
8><U
+008
9>N
A81
010
800
4361
1633
720
0R
en
wic
k G
alle
ry37
3555
9FA
LSE
1715
7335
4907
8020
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
5797
8e73
3ae
f846
02e
d9d
86c9
5fe
8ea/
5B07
4B80
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c13
5.0.
810.
810/
2789
3935
_178
2177
9612
4301
_233
6175
0346
2568
7552
_n.j
pg
FALS
EB
fPgp
KSB
CG
l15
1875
1399
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
ad54
66f5
8c16
aa9c
f210
10e
65b
18c4
7/5B
1FB
38A
/t51
.288
5-15
/e35
/278
9393
5_17
8217
7961
2430
1_23
3617
5034
6256
8755
2_n
.jp
gI c
ou
ld li
ve in
th
ese
flu
ffy
swe
et
po
tato
es
ho
ne
stly
<f0
><U
+009
F><U
+00A
4><U
+00A
4> o
h..
.th
e r
oas
ted
bla
ck p
ep
pe
r ch
icke
n a
nd
asp
arag
us
we
re g
oo
d t
oo
lol b
ut
we
kn
ow
wh
o t
he
tru
e s
tar
of
this
dis
h is
<f0
><U
+009
F><U
+009
8><U
+008
E><f
0><U
+009
F><U
+009
4><U
+00A
5>N
A81
010
800
2444
1682
30
0Th
e M
all (
Was
hin
gto
n D
C)
2364
7152
2
10FA
LSE
1715
7335
2859
3690
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
f424
b91
1982
3e47
f81f
9867
b5e
397d
a/5B
18D
5AA
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/278
9087
6_16
6981
7172
8004
2_33
1192
4903
4347
8067
2_n
.jp
gFA
LSE
BfP
go3N
DgC
H15
1875
1396
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
040f
7941
7bf5
6a09
a22b
0b7a
3fd
eb
02/5
B20
2CC
B/t
51.2
885-
15/e
35/2
7890
876_
1669
8171
7280
042_
3311
9249
0343
4780
672_
n.j
pg
Ch
eck
Us
Ou
t EV
ERY
Satu
rday
N
A10
8010
800
1919
1220
750
0Sm
ith
son
ian
s N
atio
nal
Mu
seu
m o
f N
atu
ral H
isto
ry21
4638
086
11FA
LSE
1715
7333
0836
7520
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
4cab
4b6c
5559
6dd
7ee
04f1
b22
b21
1c7/
5B1E
8E30
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/278
7808
4_14
7310
7392
7114
1_88
6637
8504
1496
6374
4_n
.jp
gFA
LSE
BfP
glq
GjV
M3
1518
7513
70h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/254
4b25
db
eb
a2ce
d54
171d
3689
a4c8
a9/5
B05
3251
/t51
.288
5-15
/e35
/278
7808
4_14
7310
7392
7114
1_88
6637
8504
1496
6374
4_n
.jp
gC
he
ck U
s O
ut
EVER
Y Sa
turd
ay
NA
1080
1080
019
1912
2075
00
Nat
ion
al M
use
um
of
Am
eri
can
His
tory
8389
99
12FA
LSE
1715
7332
8171
6390
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/6
504a
4624
7f61
2c30
9bfa
2f77
a84e
b51
/5B
1786
9A/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/2
7578
779_
4212
7125
8303
237_
7187
2142
4610
9773
824_
n.j
pg
FALS
EB
fPgl
RSB
Uo
e1518
7513
67h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/4fa
35e
2fa8
3b0e
618a
7825
9f08
b11
14c/
5B1E
15FB
/t51
.288
5-15
/e35
/275
7877
9_42
1271
2583
0323
7_71
8721
4246
1097
7382
4_n
.jp
gP
ost
wo
rko
ut
lun
ch
NA
1080
1080
062
3940
8513
00
Hir
shh
orn
Mu
seu
m a
nd
Scu
lptu
re G
ard
en
1757
70
13FA
LSE
1715
7331
1772
9170
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
aea4
4ec7
226e
690b
ac91
aefb
f9e
c436
/5B
23B
3D8/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/275
8140
7_20
2325
2321
0231
32_7
2678
3079
5140
5957
12_n
.jp
gFA
LSE
BfP
gi4j
nu
w715
1875
1347
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
9487
f480
89b
abb
6223
52f5
69aa
8d40
0/5B
0D63
F2/t
51.2
885-
15/e
35/2
7581
407_
2023
2523
2102
3132
_726
7830
7951
4059
5712
_n.j
pg
NO
NEE
D T
O T
RY,
IT C
OM
ES N
ATU
RA
LLY.
#st
ory
ofm
ylif
eN
A10
8010
801
4404
6870
60
1Th
e G
eo
rge
Was
hin
gto
n U
niv
ers
ity
2150
339
14FA
LSE
1715
7299
7539
6570
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
fecc
07e
4a27
4d1e
5bb
c045
d54
d72
9b4/
5A88
8985
/t51
.288
5-15
/e15
/c15
7.0.
405.
405/
2787
8406
_147
0271
4263
2873
_331
4647
1095
4156
0320
_n.j
pg
TRU
EB
fPf1
KC
ALH
b1518
7513
43h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/ac0
ef0
84c3
6ab
e14
19d
126a
0b2c
ff5f
a/5A
88D
8C6/
t51.
2885
-15/
e15
/278
7840
6_14
7027
1426
3287
3_33
1464
7109
5415
6032
0_n
.jp
gH
ad a
gre
at c
on
vers
atio
n a
bo
ut
#So
uth
Sud
an a
nd
ho
w I
star
ted
@ju
nu
bga
me
s w
ith
Wit
h N
ancy
. It
was
re
ally
fan
th
ou
gh a
nd
sh
e w
as m
y @
op
rah
#Fe
ed
you
rde
stin
y #a
cto
r #s
tayf
ocu
sed
#d
est
iny
#gam
ed
ev
#gam
es4
pe
ace
#h
op
e #
Pe
ace
#co
mm
un
ity
#re
fuge
es
get
the
fu
lly
Co
nve
rsio
n o
n Y
ou
Tub
e. F
oll
ow
@ju
nu
bga
me
s4
421
750
421
0707
4902
14
Lijs
t va
n b
eke
nd
e m
en
sen
uit
Was
hin
gto
n D
.C.
5762
3358
9
15FA
LSE
1715
7330
5263
1670
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
bc8
95d
118e
6961
81f7
c965
91ac
1900
2/5B
1143
FE/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
0.13
4.10
80.1
080/
2815
0890
_330
7703
7410
0855
_391
2232
3191
8794
3424
_n.j
pg
FALS
EB
fPgh
77ge
9K151
8751
339
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
372f
ee
160a
63e
3ad
d9d
3c64
98e
15e
00/5
B06
AA
5C/t
51.2
885-
15/e
35/2
8150
890_
3307
7037
4100
855_
3912
2323
1918
7943
424_
n.j
pg
<f0>
<U+0
09F>
<U+0
090>
<U+0
0BE>
<f0>
<U+0
09F>
<U+0
090>
<U+0
0BE>
<f0>
<U+0
09F>
<U+0
090>
<U+0
0BE>
<f0>
<U+0
09F>
<U+0
090>
<U+0
0BE>
<f0
><U
+009
F><U
+009
9><U
+008
C><
f0><
U+0
09F>
<U+0
08F>
<U+0
0BD
><f0
><U
+009
F><U
+009
9><U
+008
C><
f0><
U+0
09F>
<U+0
08F>
<U+0
0BD
><f0
><U
+009
F><U
+009
9><U
+008
C><
f0><
U+0
09F>
<U+0
08F>
<U+0
0BD
>N
A13
4910
802
3929
0737
460
2N
atio
nal
Po
rtra
it G
alle
ry, S
mit
hso
nia
n In
stit
uti
on
2165
5036
3
16FA
LSE
1715
7330
4406
8840
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
21d
3892
b93
9198
186a
dc2
b3a
1c89
4d7/
5B09
6F5A
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c0.
135.
1080
.108
0/27
8782
50_1
6437
5166
9041
827_
4684
7905
1074
3715
84_n
.jp
gFA
LSE
BfP
ghz9
H39
O1518
7513
38h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/a6e
57d
530c
8490
86b
f1d
cd1e
e57
7097
5/5B
0A8C
88/t
51.2
885-
15/e
35/2
7878
250_
1643
7516
6904
1827
_468
4790
5107
4371
584_
n.j
pg
#Lin
geri
eP
arty
<f0
><U
+009
F><U
+009
8><U
+008
8>N
A13
5010
803
1700
3780
910
3G
eo
rge
tow
n, D
C10
0999
7177
17FA
LSE
1715
7330
0677
3110
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
9b11
a058
757f
b3b
c520
8b57
9ecd
f88a
/5B
031A
F9/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
0.13
4.10
80.1
080/
2758
0427
_142
9938
1137
9476
0_62
6025
1860
0037
1712
0_n
.jp
gFA
LSE
BfP
ghR
OH
-bV
1518
7513
34h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/4d
54d
ee
37f2
1b05
3230
e49
e68
8b1c
5f3/
5B13
DD
26/t
51.2
885-
15/e
35/2
7580
427_
1429
9381
1379
4760
_626
0251
8600
0371
7120
_n.j
pg
Trav
ell
ing
to U
nio
n M
arke
t N
A13
4910
800
7107
3720
430
0N
atio
nal
Wo
rld
War
II M
em
ori
al20
9182
3
18FA
LSE
1715
7326
8374
2550
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
adf7
37c3
7544
4d25
9b4c
f5b
c1f5
4107
/5B
26C
BA
5/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
135.
0.81
0.81
0/27
5805
46_1
4539
8566
2716
44_2
0907
4757
2401
4714
88_n
.jp
gFA
LSE
BfP
gckY
AK
UY
1518
7512
96h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/425
05a7
838d
f127
4b19
7c47
fcd
4a9a
23/5
B1D
CEA
F/t5
1.28
85-1
5/e
35/2
7580
546_
1453
9856
6271
644_
2090
7475
7240
1471
488_
n.j
pg
<U+3
0C8>
<U+3
0EC
><U
+30B
8><U
+30E
7><U
+592
7><U
+597
D><
U+3
04D
>!<U
+6C
38><
U+9
060>
<U+3
06B
><U
+304
4><U
+308
C><
U+3
07E>
<U+3
059>
<U+3
06D
><U
+7B
11>
<U+7
25B
><U
+4E7
3><U
+306
8><U
+30D
7><U
+30E
D><
U+3
0B7>
<U+3
0E5>
<U+3
0FC
><U
+30C
8><U
+306
0><U
+305
1><U
+8C
B7>
<U+3
046>
<U+3
064>
<U+3
082>
<U+3
08A
><U
+306
0><U
+306
3><U
+305
F><U
+306
E><U
+306
B><
U+3
044>
<U+3
08D
><U
+304
4><U
+308
D><
U+8
A66
><U
+305
7><U
+306
6><U
+307
F><U
+305
F><U
+304
F><U
+306
6><U
+8C
B7>
<U+3
063>
<U+3
061>
<U+3
083>
<U+3
046>
<U+2
728>
<U
+305
F><U
+304
F><U
+305
5><U
+309
3><U
+976
2><U
+767
D><
U+3
044>
<U+3
082>
<U+3
06E>
<U+7
67A
><U
+898
B><
U+3
057>
<U+3
05F>
<U+3
06E>
<U+3
067>
<U+3
07E>
<U+3
05F>
<U+8
F09>
<U+3
05B
><U
+306
6><U
+304
4><U
+304
D><
U+3
07E>
<U+3
059>
NA
810
1080
847
5174
7131
18
Ne
wse
um
2613
1
19FA
LSE
1715
7323
9566
9630
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
3a3c
83c4
5a35
5bb
5252
a031
fe4c
4534
/5B
1F69
54/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
0.13
5.10
80.1
080/
2787
9225
_178
2405
0218
0421
8_49
9259
4736
4904
7961
6_n
.jp
gFA
LSE
BfP
gYYF
hN
KM
1518
7512
61h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/cc3
dd
83ae
4911
4c3c
3f3a
09d
fd08
b43
d/5
B20
6C57
/t51
.288
5-15
/e35
/278
7922
5_17
8240
5021
8042
18_4
9925
9473
6490
4796
16_n
.jp
gH
ad v
ayca
y. N
ee
d r
un
s. A
nd
N
A13
5010
803
6577
2976
840
3D
up
on
t C
ircl
e59
0718
213
20FA
LSE
1715
7319
6233
9060
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
40d
922a
a944
82d
1bab
d08
6b89
fa14
e2/
5B16
F0A
B/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/2
7578
967_
1853
6310
8894
430_
7987
9199
0097
6873
472_
n.j
pg
FALS
EB
fPgS
EhA
P6u15
1875
1210
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/f
96d
379f
cc90
951d
91f4
5659
b47
df2
74/5
B0F
59C
A/t
51.2
885-
15/e
35/2
7578
967_
1853
6310
8894
430_
7987
9199
0097
6873
472_
n.j
pg
It's
no
t e
ven
th
at w
arm
an
d I'
m a
ll e
xcit
ed
! #h
app
y #a
llsm
ile
s #d
cnig
hts
#il
ltak
eit
<f0
><U
+009
F><U
+009
8><U
+008
A><
f0><
U+0
09F>
<U+0
098>
<U+0
08A
><U
+260
0><U
+FE0
F><f
0><U
+009
F><U
+008
C><
U+0
0B4>
<f0>
<U+0
09F>
<U+0
091>
<U+0
0A0>
<f0>
<U+0
09F>
<U+0
095>
<U+0
0B6>
<f0
><U
+009
F><U
+009
4><U
+00A
5>N
A10
8010
800
6835
4931
50
0W
alte
r E.
Was
hin
gto
n C
on
ven
tio
n C
en
ter
2161
4
21FA
LSE
1715
7319
2678
8380
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/0
dc8
c10a
bf2
6975
7103
dcd
b5e
afe
4e6f
/5B
0EC
80B
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c0.
134.
1080
.108
0/27
5786
71_1
5582
8578
5481
15_4
7184
5570
1360
3450
88_n
.jp
gFA
LSE
BfP
gRja
BJq
_1518
7512
05h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/be
40c7
8858
6b84
bd
0712
2c85
f059
c76b
/5B
0324
A9/
t51.
2885
-15/
e35
/275
7867
1_15
5828
5785
4811
5_47
1845
5701
3603
4508
8_n
.jp
gTh
e #
LBD
ge
ts a
#b
oss
lad
y N
A13
4910
808
3901
9971
630
8Ti
dal
Bas
in34
2980
053
22FA
LSE
1715
7317
2968
1450
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/7
9c93
755e
ab15
30d
56a0
9248
6acd
52ac
/5B
1D69
31/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
0.13
4.10
80.1
080/
2758
0806
_955
5956
1795
2292
_110
9927
3400
8062
7712
_n.j
pg
FALS
EB
fPgO
r1h
5VQ
1518
7511
82h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/790
2774
efc
c69f
3ae
99f6
1260
de
d5d
89/5
B25
7093
/t51
.288
5-15
/e35
/275
8080
6_95
5595
6179
5229
2_11
0992
7340
0806
2771
2_n
.jp
gif
yo
u w
ann
a co
me
th
ru, t
he
re's
a s
eat
just
fo
r yo
uN
A13
4910
8011
2236
9224
011
Re
fle
ctin
g P
oo
l21
5073
23FA
LSE
1715
7316
2858
2960
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
2bb
de
1ce
2459
dfa
7139
3652
7ca6
8e4f
/5B
218B
F6/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/2
7577
130_
4450
7028
2582
638_
6151
4230
7525
9777
024_
n.j
pg
FALS
EB
fPgN
Nrl
wu
B15
1875
1170
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/0
b86
88b
cc9d
e83
adfe
30d
0a3d
b21
b1a
3/5B
0168
97/t
51.2
885-
15/e
35/2
7577
130_
4450
7028
2582
638_
6151
4230
7525
9777
024_
n.j
pg
He
re w
/ Fa
shio
n d
esi
gne
r/st
ylis
t K
rist
i Lo
ve @
kris
tilo
vest
yle
Th
e #
ho
war
du
niv
ers
ity
@h
ow
ard
1867
alu
m w
rite
s th
e @
ifu
we
rem
ybo
yfri
en
d b
log
and
is o
ne
of
the
be
st s
tyli
sts
in t
he
bu
sin
ess
! #b
lack
girl
mag
ic #
ho
war
du
#h
ub
iso
np
rid
e #
hu
ukn
ow
NA
1080
1080
322
6050
3050
03
Mad
ame
Tu
ssau
ds
DC
6834
22
24FA
LSE
1715
7312
5513
7790
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
47fc
c8d
4128
6db
9278
9d1d
f1f1
04e
43/5
B26
88EB
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/278
9152
4_23
9868
5398
8829
2_29
0513
8706
7655
7824
0_n
.jp
gFA
LSE
BfP
gHx4
hf7
l1518
7511
25h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/c64
7847
b19
f084
2397
4923
3f3f
3a3d
ac/5
B15
878A
/t51
.288
5-15
/e35
/278
9152
4_23
9868
5398
8829
2_29
0513
8706
7655
7824
0_n
.jp
gA
ne
w D
C m
ura
l, s
pie
d it
to
day
in N
OM
A!
NA
1080
1080
053
5313
9382
00
Le D
iplo
mat
e75
3679
54
25FA
LSE
1715
7311
7700
6010
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
bb
aec4
8686
e6d
101b
c924
f0f0
6eae
27/5
B1D
AB
CF/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c0.
0.10
79.1
079/
2789
2878
_198
9111
3180
1226
0_10
7532
3458
5916
5388
8_n
.jp
gFA
LSE
BfP
gGp
HgZ
x_1518
7511
16h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/8c1
7c4e
890f
c836
aa4e
e23
fc9b
db
e52
f/5B
0AC
704/
t51.
2885
-15/
e35
/278
9287
8_19
8911
1318
0122
60_1
0753
2345
8591
6538
88_n
.jp
gN
AN
A10
7910
801
5110
557
01
Inte
rnat
ion
al S
py
Mu
seu
m2178
3
26FA
LSE
1715
7307
7295
3710
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
546e
614d
c19e
f1e
d9d
d45
5b4a
a8e
196/
5B01
64B
1/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/2
7580
950_
1555
6780
6788
4089
_883
5595
5243
5145
9328
_n.j
pg
FALS
EB
fPgA
w0H
OA
J15
1875
1068
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/4
356f
5f8f
e9c
c93c
1d79
2d2f
f0e
89d
19/5
B23
D69
B/t
51.2
885-
15/e
35/2
7580
950_
1555
6780
6788
4089
_883
5595
5243
5145
9328
_n.j
pg
NO
T TO
DA
Y, S
ATA
N.
NA
1080
1080
1532
2803
9039
015
Nat
ion
al M
use
um
of
Wo
me
n in
th
e A
rts
2187
2346
4
27FA
LSE
1715
7307
1828
4650
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/6
6309
f999
a406
935a
ce1e
a2e
3dd
50a2
6/5B
10F2
71/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
135.
0.81
0.81
0/27
5820
99_1
5449
4756
2292
648_
1856
7696
2969
9964
928_
n.j
pg
FALS
EB
fPf_
95lS
SH15
1875
1061
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
50a0
092b
2166
5445
1d26
91b
be
5b54
54/5
B0C
BC
F4/t
51.2
885-
15/e
35/2
7582
099_
1544
9475
6229
2648
_185
6769
6296
9996
4928
_n.j
pg
No
t a
goo
d r
un
, bu
t a
pre
tty
nig
ht.
70
de
gre
es
in F
eb
ruar
y fe
els
pre
tty
goo
d!
#dc
#was
hin
gto
nd
c #
nig
htr
un
#fe
els
like
spri
ng
#nat
ion
alm
all #
nig
ht
#lig
hts
#ig
dc
#byt
hin
gs #
dcf
ocu
sed
#in
stan
tdc
#exp
ose
dd
cN
A81
010
803
6407
140
3W
ash
ingt
on
, DC
2128
9705
9
28FA
LSE
1715
7305
2205
0600
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/7
b42
e34
6d44
b1b
9ed
702b
4ae
3c8a
0e02
/5B
22C
620/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/278
7820
2_19
1903
8795
0742
82_8
7509
3389
3516
8860
16_n
.jp
gFA
LSE
BfP
f9H
JHz5
o1518
7510
38h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/f58
201d
1602
004e
03e
f45a
46f1
bfc
7f8/
5B1F
A60
A/t
51.2
885-
15/e
35/2
7878
202_
1919
0387
9507
4282
_875
0933
8935
1688
6016
_n.j
pg
Foll
ow
me
at
@co
rin
ne
the
do
gmo
m o
r yo
u'r
e a
fak
e f
rie
nd
<U
+276
4><U
+FE0
F>N
A10
8010
807
5138
3746
61
7M
arti
n L
uth
er
Kin
g, J
r. N
atio
nal
Me
mo
rial
1420
252
29FA
LSE
1715
7304
1943
2800
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
8e75
40d
7137
8242
538b
1bb
be
fdd
342b
/5B
2602
2E/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
180.
0.71
9.71
9/27
5786
80_9
8421
9838
3826
03_4
7840
9228
4931
0146
56_n
.jp
gFA
LSE
BfP
f7n
kn9O
d1518
7510
26h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/219
53a8
553e
7714
5622
b01
b56
8254
4f2/
5B1A
BFC
9/t5
1.28
85-1
5/e
35/2
7578
680_
9842
1983
8382
603_
4784
0922
8493
1014
656_
n.j
pg
Ru
th E
. Car
ter
NA
719
1080
893
0631
721
08
Smit
hso
nia
ns
Fre
er
and
Sac
kle
r G
alle
rie
s11
1955
31
30FA
LSE
1715
7302
0655
3110
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
e1d
9977
6f53
7803
4f7f
ad0d
400e
ee
11/5
B21
383A
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/275
8162
1_18
9650
9116
3158
2_41
8317
4524
2603
1923
2_n
.jp
gFA
LSE
BfP
f4h
UA
Xn
415
1875
1000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
30b
9eb
fa46
93ff
d46
1427
8b34
5bf7
39/5
B08
1C5B
/t51
.288
5-15
/e35
/275
8162
1_18
9650
9116
3158
2_41
8317
4524
2603
1923
2_n
.jp
gTh
is is
ou
r ch
ild
ren
Am
eri
ca. T
he
re is
a p
rob
lem
an
d s
om
eth
ing
ne
ed
s to
giv
e. #
amo
vert
ho
ugh
tsan
dp
raye
rs #
lets
getu
nco
mfo
rtab
le #
con
gre
ss #
po
tus
NA
1080
1080
120
9523
7453
01
Vas
hin
gto
n, D
C10
2427
2801
31FA
LSE
1715
7301
5274
8960
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
181d
c05c
891d
0da7
7d7e
d09
1887
55f8
/5B
1A38
7B/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/2
7578
804_
1583
6542
2839
5841
_323
5027
5469
7419
1616
_n.j
pg
FALS
EB
fPf3
vNB
1MQ
1518
7509
94h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/bab
fe2a
a2ce
b5a
d56
cc06
7e1e
1005
7e3/
5B04
9B51
/t51
.288
5-15
/e35
/275
7880
4_15
8365
4228
3958
41_3
2350
2754
6974
1916
16_n
.jp
gC
on
sve
rati
ves
wil
l ne
ver
cham
pio
n m
en
tal h
eal
th. T
he
y ch
amp
ion
gu
ns.
<f0
><U
+009
F><U
+009
4><U
+00A
B><
f0><
U+0
09F>
<U+0
094>
<U+0
0AB
><f0
><U
+009
F><U
+009
4><U
+00A
B>
NA
1080
1080
1045
8836
1131
010
W W
ash
ingt
on
D.C
. Ho
tel
1334
59
32FA
LSE
1715
7299
7511
9910
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/7
16e
95e
ace
6536
e16
7810
d19
68c6
ce37
/5B
253A
B1/
t51.
2885
-15/
e35
/275
7589
3_18
0879
4619
4193
36_7
4078
3004
7337
2180
48_n
.jp
gFA
LSE
BfP
f1Jx
gy7l1
5187
5097
3h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/716
e95
eac
e65
36e
1678
10d
1968
c6ce
37/5
B25
3AB
1/t5
1.28
85-1
5/e
35/2
7575
893_
1808
7946
1941
9336
_740
7830
0473
3721
8048
_n.j
pg
Act
ivo
Co
n e
l #Fl
ow
LeP
ew
<U+2
03C
><U
+FE0
F><f
0><U
+009
F><U
+009
2><U
+00A
F><f
0><U
+009
F><U
+009
4><U
+00A
5>N
A48
048
02
2811
4602
60
2Fo
un
din
g Fa
rme
rs29
5793
33FA
LSE
1715
7298
1271
6550
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
63b
1150
efd
91b
0467
1939
a7e
98b
858c
/5B
05FD
BE/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/275
8214
0_16
4179
0575
8870
44_9
2182
2995
1684
1492
48_n
.jp
gFA
LSE
BfP
fyyh
hIg
01518
7509
53h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/a20
3f7b
e11
0434
64f6
e75
2a48
eab
4cd
2/5B
0125
94/t
51.2
885-
15/e
35/2
7582
140_
1641
7905
7588
7044
_921
8229
9516
8414
9248
_n.j
pg
Ro
man
vib
es
thro
wb
ack
NA
1080
1080
1021
3383
616
210
Cit
yCe
nte
rDC
9409
4015
4
34FA
LSE
1715
7295
6910
3050
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/6
3271
8f33
07fe
8be
f82e
be
be
5551
bc0
1/5B
0F35
AE/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c0.
135.
1080
.108
0/27
5772
25_1
7608
4838
4222
142_
7510
1320
7951
9072
256_
n.j
pg
FALS
EB
fPfv
Pp
BX
fs151
8750
924
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
b2b
7bd
41b
4248
527a
6824
fa61
cb47
05/5
B21
8BA
D/t
51.2
885-
15/e
35/2
7577
225_
1760
8483
8422
2142
_751
0132
0795
1907
2256
_n.j
pg
bo
ots
wit
h t
he
fu
rN
A13
5010
8048
2088
8638
348
Vie
tnam
Ve
tera
ns
Me
mo
rial
1331
62
35FA
LSE
1715
7295
2170
7220
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
c3d
0582
0660
d37
78af
e9c
df1
70e
253e
/5B
03A
E37/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c0.
108.
1080
.108
0/27
8786
32_3
7816
1569
3228
84_5
7395
5702
7496
0670
72_n
.jp
gFA
LSE
BfP
fujg
Ap
OG15
1875
0919
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/4
a7fd
76b
e1b
1375
91b
6f1f
a7e
24ff
b40
/5B
249C
B5/
t51.
2885
-15/
e35
/278
7863
2_37
8161
5693
2288
4_57
3955
7027
4960
6707
2_n
.jp
gLe
o h
as m
ade
a n
ew
be
st f
rie
nd
, th
e o
ne
an
d o
nly
Jar
ell
. #h
ou
segu
est
#le
oth
ech
igi #
chig
i #d
cdo
gs #
dcd
ogm
om
s #w
eth
ed
ogs
dc
#do
gso
fin
stag
ram
#in
stad
og
#ch
ihu
ahu
aso
fin
stgr
am #
frie
nd
sN
A12
9710
807
1497
6885
731
7D
AR
Co
nst
itu
tio
n H
all
5658
61
36FA
LSE
1715
7294
7953
9000
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
7aff
64cf
3908
c161
2e66
760d
459c
906/
5B18
2DB
D/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
180.
0.71
9.71
9/25
0388
83_2
6218
7704
3169
23_2
4599
4461
0433
0731
52_n
.jp
gFA
LSE
BfP
ft8O
low
61518
7509
14h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/044
3e3d
8fa4
028e
6ca4
8806
9c08
5314
9/5B
2368
5A/t
51.2
885-
15/e
35/2
5038
883_
2621
8770
4316
923_
2459
9446
1043
3073
152_
n.j
pg
“Dam
e u
n 'e
po
co lo
ve w
ith
N
A71
910
8020
1527
6931
330
20Th
e H
amil
ton
7025
540
37FA
LSE
1715
7294
4763
6080
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
450c
12e
ee
5b38
f1f4
872e
2d4f
20cb
12/5
B04
60A
4/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/2
7579
033_
1058
0241
0433
5485
_562
7873
0988
1760
1536
_n.j
pg
FALS
EB
fPft
eh
BrU
D1518
7509
10h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/7a8
8341
332e
6a39
fe1e
c54b
d18
b88
025/
5B19
BA
8E/t
51.2
885-
15/e
35/2
7579
033_
1058
0241
0433
5485
_562
7873
0988
1760
1536
_n.j
pg
Had
a lo
vely
eve
nin
g w
ith
my
big
bir
d. <
U+2
764>
<U+F
E0F>
<f0>
<U+0
09F>
<U+0
090>
<U+0
0A5>
NA
1080
1080
032
2203
040
0U
nit
ed
Sta
tes
Ho
loca
ust
Me
mo
rial
Mu
seu
m21
5230
297
38FA
LSE
1715
7293
3336
7870
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
9219
eb
32fd
e86
2af5
58e
b76
54d
8272
0/5B
08C
B1E
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/275
8112
6_19
6037
8043
1430
2_24
2188
3439
1846
0928
_n.j
pg
FALS
EB
fPfr
0GH
Djh1
5187
5089
6h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/ec5
2111
1a43
a857
2892
10e
6ca7
e85
7e4/
5B05
DD
8B/t
51.2
885-
15/e
35/2
7581
126_
1960
3780
4314
302_
2421
8834
3918
4609
28_n
.jp
gI’
m h
ere
, in
my
pla
ce, f
inal
ly <
U+2
764>
<U+F
E0F>
#w
ash
ingt
on
dc
#was
hin
gto
nm
on
um
en
t #a
shb
urn
virg
inia
#so
uls
iste
rtim
eN
A10
8010
802
2026
8692
40
2W
orl
d W
ar II
Me
mo
rial
4644
9657
0
39FA
LSE
1715
7289
8226
2310
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/0
156a
b1b
9bb
2fc1
e25
a324
ae6a
d9d
72e
/5B
2434
56/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/2
8151
281_
5860
8067
1730
528_
2002
6587
3805
3316
608_
n.j
pg
FALS
EB
fPfm
tGlm
0E1518
7508
54h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/a5d
e1f
5b95
1c8b
e79
ccf1
e80
cb91
d5b
c/5B
126B
37/t
51.2
885-
15/e
35/2
8151
281_
5860
8067
1730
528_
2002
6587
3805
3316
608_
n.j
pg
Do
nt
me
an t
o b
reak
yo
ur
he
art
bu
t b
aby
that
wh
at t
hu
gs d
oN
A10
8010
8012
5974
3200
671
12W
arn
er
The
atre
8440
5
40FA
LSE
1715
7286
5544
2610
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
4b76
e7c
7755
671d
c3a7
52a8
b2d
09f5
9/5B
1B5B
93/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
0.13
5.10
80.1
080/
2757
6851
_204
5104
0024
3369
6_40
1813
1644
1782
3539
2_n
.jp
gFA
LSE
BfP
fh8u
nZT
61518
7508
15h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/544
b7c
f903
7e25
bc6
5952
9cb
8b5f
36e
a/5B
1C54
90/t
51.2
885-
15/e
35/2
7576
851_
2045
1040
0243
3696
_401
8131
6441
7823
5392
_n.j
pg
NA
NA
1350
1080
2331
4298
8762
023
Smit
hso
nia
n A
me
rica
n A
rt M
use
um
an
d t
he
Re
nw
ick
Gal
lery
2139
3359
2
41FA
LSE
1715
7286
0252
5650
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
6a81
4c93
54fd
f211
2547
34f8
e2e
f86a
/5B
15D
93D
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c0.
135.
1080
.108
0/28
1546
64_2
2555
9704
6784
18_5
9770
2808
4279
2796
16_n
.jp
gFA
LSE
BfP
fhLc
hM
zN1518
7508
09h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/5c2
3305
a7a6
d1c
51d
a0d
6e5c
2b67
1ecb
/5B
073C
EF/t
51.2
885-
15/e
35/2
8154
664_
2255
5970
4678
418_
5977
0280
8427
9279
616_
n.j
pg
Swe
et
NA
1350
1080
280
1423
2494
928
0N
atio
nal
Ch
rist
mas
Tre
e26
5107
589
42FA
LSE
1715
7285
0003
3810
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
b0b
b0d
da3
de
b3c
2ac9
c401
b3d
e99
dd
e/5
B06
6924
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c0.
31.8
08.8
08/2
7880
920_
3961
1533
7495
260_
3695
2043
4466
1016
576_
n.j
pg
FALS
EB
fPff
r_h
5Ok15
1875
0797
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
70b
41c8
ea6
695e
5585
893e
0013
1662
f/5B
0135
66/t
51.2
885-
15/e
35/2
7880
920_
3961
1533
7495
260_
3695
2043
4466
1016
576_
n.j
pg
Wh
o w
ants
to
ge
t cr
azy
wit
h m
e in
May
?! #
star
cou
rse
#50
mil
er
#cla
ss00
1 #g
oru
ck #
em
bra
ceth
esu
ck #
ruck
ing
NA
1162
1080
240
2854
653
2W
ash
ingt
on
, D.C
2939
8462
8
43FA
LSE
1715
7284
6238
6570
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
8935
59cf
4524
4763
77fc
d83
3df5
8973
/5B
23C
A05
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/278
9105
3_15
5452
6685
9321
8_57
2383
5759
2705
5974
4_n
.jp
gFA
LSE
BfP
ffI7
lDja
1518
7507
92h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/673
eb
63b
36e
967d
2bd
fa30
20d
7019
252/
5B24
5864
/t51
.288
5-15
/e35
/278
9105
3_15
5452
6685
9321
8_57
2383
5759
2705
5974
4_n
.jp
g#t
hro
wb
ackt
hu
rsd
ay B
ack
in t
he
day
wh
en
I h
ad lu
nch
ed
ou
t o
ff t
he
wat
er!
!! <
f0><
U+0
09F>
<U+0
0A4>
<U+0
0A3>
<f0>
<U+0
09F>
<U+0
0A4>
<U+0
0A3>
<f0>
<U+0
09F>
<U+0
0A4>
<U+0
0A3>
<f0>
<U+0
09F>
<U+0
0A4>
<U+0
0A3>
<f0>
<U+0
09F>
<U+0
0A4>
<U+0
0A3>
jk I
ain
't s
mo
ke n
o w
ate
r lm
aoN
A10
8010
807
4739
9013
27
Ab
rah
am L
inco
ln M
em
ori
al21
2895
774
44FA
LSE
1715
7283
5658
8390
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/f
dfe
6895
40e
7ee
e03
3648
63b
eb
f769
ab/5
B12
836F
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c0.
135.
1080
.108
0/27
5810
94_1
5347
5115
4614
14_9
3343
3463
0070
8454
4_n
.jp
gFA
LSE
BfP
fdm
ZhC
e8
1518
7507
80h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/1f3
892d
dc8
6522
4bd
c55a
7fa3
0816
ae6/
5B02
8C79
/t51
.288
5-15
/e35
/275
8109
4_15
3475
1154
6141
4_93
3433
4630
0708
4544
_n.j
pg
Wit
h y
o N
atu
ral h
air,
yo
ur
soft
ski
n a
nd
yo
ur
big
ass
in t
hat
su
n d
ress
<f0
><U
+009
F><U
+009
8><U
+008
D><
f0><
U+0
09F>
<U+0
0A4>
<U+0
0AF>
NA
1350
1080
3133
7085
140
31A
we
som
e C
on
8562
2440
45FA
LSE
1715
7281
8729
0080
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
82e
24f2
6a4c
a78e
2361
a11f
3be
404c
1/5B
0555
B2/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c0.
113.
1080
.108
0/27
5768
00_4
2490
9381
2779
91_8
2098
9827
6284
0064
0_n
.jp
gFA
LSE
BfP
fbIu
jPH
M1518
7507
59h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/24a
27c9
cb2c
ab4d
e7e
f4fd
c51b
7163
ba/
5B00
F2C
6/t5
1.28
85-1
5/e
35/2
7576
800_
4249
0938
1277
991_
8209
8982
7628
4006
40_n
.jp
gN
ot
you
r av
era
ge o
r ty
pic
al <
f0><
U+0
09F>
<U+0
091>
<U+0
0BC
><f0
><U
+009
F><U
+008
F><U
+00B
D>
NA
1307
1080
4051
7858
021
40Th
e P
ark
at 1
4th
4393
92
46FA
LSE
1715
7273
4694
3860
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
891c
bf4
ffd
43cd
1f18
8aad
d52
84d
be
3/5A
88D
F56/
t51.
2885
-15/
e15
/275
8145
5_14
9276
9424
5109
0_37
1417
1134
9885
1737
6_n
.jp
gTR
UE
BfP
fO6G
AzE
c15
1875
0694
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
891c
bf4
ffd
43cd
1f18
8aad
d52
84d
be
3/5A
88D
F56/
t51.
2885
-15/
e15
/275
8145
5_14
9276
9424
5109
0_37
1417
1134
9885
1737
6_n
.jp
gM
ore
Od
d L
ou
ise
on
th
e r
ise
#th
yse
lfco
lle
ctio
n15
640
640
227
3684
480
2Tr
um
p In
tern
atio
nal
Ho
tel W
ash
ingt
on
, D.C
.10
2493
6088
47FA
LSE
1715
7266
9035
9130
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
5260
d38
a72b
920b
656b
09e
491a
2834
9/5A
8889
17/t
51.2
885-
15/s
640x
640/
e15
/c0.
89.7
20.7
20/2
7581
061_
1583
8222
1624
105_
7696
5880
0438
6668
544_
n.j
pg
TRU
EB
fPfF
Wm
gzn
R15
1875
0663
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
15ad
5f5e
bd
ccd
c66d
9039
e24
a58f
12e
/5A
8890
36/t
51.2
885-
15/e
15/2
7581
061_
1583
8222
1624
105_
7696
5880
0438
6668
544_
n.j
pg
shh
hh
... i
’m h
idin
g <f
0><U
+009
F><U
+00A
4><U
+00A
B><
f0><
U+0
09F>
<U+0
099>
<U+0
088>
<f0>
<U+0
09F>
<U+0
098>
<U+0
086>
@th
ee
lle
nsh
ow
#b
uzz
fee
d #
inst
ago
od
#in
stad
og
#do
goft
he
day
#sh
arp
ei #
shar
pe
ilo
ve #
cute
ne
sso
verl
oad
#fu
nn
yvid
eo
s #h
ide
and
see
k #s
har
pe
ioft
he
day
#to
psh
arp
ei #
we
ekl
yflu
ff #
shar
pe
iso
fin
stag
ram
#d
c #d
ogs
ofd
c18
936
750
1359
6177
6691
013
Old
Eb
bit
t G
rill
4911
1800
48FA
LSE
1715
7260
1081
6060
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/f
32c1
0a83
f77d
971e
13e
d88
0bf6
9dc5
3/5A
886A
44/t
51.2
885-
15/s
640x
640/
e15
/278
9200
9_20
4579
8809
0357
51_8
7114
8288
8031
3057
28_n
.jp
gTR
UE
BfP
e7d
ulv
Fl151
8750
661
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/7
bd
1499
36fe
715c
a942
e91
b87
fd91
1f7/
5A88
7C33
/t51
.288
5-15
/e15
/278
9200
9_20
4579
8809
0357
51_8
7114
8288
8031
3057
28_n
.jp
g..
.Fle
xin
’... *
Co
min
g So
on
* 30
750
750
1521
3250
4969
315
Was
hin
gto
n M
on
um
en
t21
7010
719
49FA
LSE
1715
7273
1680
4920
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/7
2384
f4af
f5a3
a015
0939
ee
72b
ed
3317
/5B
15A
C57
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c13
5.0.
809.
809/
2787
9429
_364
8188
5398
5012
_893
3569
5051
0423
2448
_n.j
pg
FALS
EB
fPfO
eB
l2Q
s1518
7506
56h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/b85
7ccb
8143
e25
474c
400e
455f
1461
79/5
B16
83F0
/t51
.288
5-15
/e35
/278
7942
9_36
4818
8539
8501
2_89
3356
9505
1042
3244
8_n
.jp
gTh
nx
Dal
pin
o.
NA
809
1080
3032
5987
0568
130
Mar
tin
Lu
the
r K
ing,
Jr.
Me
mo
rial
2204
1653
9
50FA
LSE
1715
7270
6450
0600
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
0153
bf5
73b
ecc
a9cc
a697
c190
aa0e
c3/5
B02
C54
5/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/2
7894
426_
1394
3760
3536
997_
8116
4150
5750
1356
032_
n.j
pg
FALS
EB
fPfK
zDFR
P715
1875
0626
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
ba0
ffd
9bb
9391
7eb
13f6
2e0c
bac
c22b
/5B
1BA
924/
t51.
2885
-15/
e35
/278
9442
6_13
9437
6035
3699
7_81
1641
5057
5013
5603
2_n
.jp
g#T
BT
<f0>
<U+0
09F>
<U+0
092>
<U+0
097>
Mar
ch 2
015
NA
1080
1080
1417
4509
918
014
Wo
rld
Ban
k81
794
51FA
LSE
1715
6896
6518
3700
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
658a
596e
9db
5264
672a
2678
50d
5a40
4/5B
1572
38/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/2
8151
637_
1731
7230
9993
010_
8191
6926
9160
0310
272_
n.j
pg
FALS
EB
fPW
qkO
FXLB
1518
7461
67h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/529
fc39
2e95
5cb
cd13
2196
6ace
0d4d
78/5
B1C
0159
/t51
.288
5-15
/e35
/281
5163
7_17
3172
3099
9301
0_81
9169
2691
6003
1027
2_n
.jp
gM
ee
tin
g th
e P
resi
de
nt
in P
ort
lan
d, O
rego
n, 2
012.
I’m
co
min
g b
ack
to P
ort
lan
d o
n M
ay 1
fo
r tw
o e
ven
ts a
t R
evo
luti
on
Hal
l sp
on
sore
d b
y P
ow
ell
’s B
oo
ks. T
icke
ts g
o o
n s
ale
to
mo
rro
w a
t 10
AM
PST
. Dir
ect
lin
k o
n m
y w
eb
site
.N
A10
8010
8025
493
4303
2581
9719
725
493
Was
hin
gto
n, D
istr
ict
of
Co
lum
bia
2134
8018
0
52FA
LSE
1715
6789
5299
0570
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/f
ace
d08
548b
8834
3be
0093
6790
b1e
40d
/5B
12D
476/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/278
7810
0_14
2996
0197
2816
8_22
5249
0955
5275
5302
4_n
.jp
gFA
LSE
BfP
UO
rtn
hvx15
1874
4890
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
3267
8055
4e6d
3103
2974
0b8d
09a1
c48/
5B0B
A21
7/t5
1.28
85-1
5/e
35/2
7878
100_
1429
9601
9728
168_
2252
4909
5552
7553
024_
n.j
pg
18: C
an I
bu
y b
ee
r?
NA
1080
1080
749
4933
0611
0329
749
The
Ob
ama
Wh
ite
Ho
use
2259
3156
5
53FA
LSE
1715
6700
4312
2260
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
eca
67e
e5b
aa3f
8a27
874d
6f1c
5e92
64/5
B15
0AA
1/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/2
7577
236_
2016
0041
4867
1483
_506
8229
3682
7107
7376
_n.j
pg
FALS
EB
fPSN
Bw
FOia
1518
7438
28h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/012
910b
a466
ee
2e64
ec6
5861
e19
0e29
e/5
B08
8F8B
/t51
.288
5-15
/e35
/275
7723
6_20
1600
4148
6714
83_5
0682
2936
8271
0773
76_n
.jp
g#T
BT
#PO
TUS4
4 @
bar
acko
bam
a d
uri
ng
an e
mo
tio
nal
sp
ee
ch o
n g
un
co
ntr
ol!
“Ev
ery
tim
e I
thin
k ab
ou
t th
ose
kid
s it
ge
ts m
e m
ad”
- B
arac
k O
bam
a (r
efe
ren
cin
g Sa
nd
y H
oo
k) <
f0><
U+0
09F>
<U+0
099>
<U+0
08F>
<f0>
<U+0
09F>
<U+0
08F>
<U+0
0BD
><f0
><U
+009
F><U
+009
9><U
+008
F><f
0><U
+009
F><U
+008
F><U
+00B
D>
#44
#cla
ssac
t #p
rote
cto
urk
ids
#pro
tect
ou
rte
ach
ers
NA
1080
1080
3751
7968
12
37Li
nco
ln M
em
ori
al15
712
54FA
LSE
1715
6676
5322
3310
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
64b
a59d
3f6e
5e4c
b55
254e
04d
abff
ba/
5B15
32B
4/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
0.11
7.93
7.93
7/27
5822
35_1
5106
2625
6075
68_7
9371
0245
3520
9902
08_n
.jp
gFA
LSE
BfP
Rq
P_A
P7E15
1874
3543
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/0
2811
c4cb
1d2c
9888
da0
c6a1
702c
8506
/5B
2026
D5/
t51.
2885
-15/
e35
/275
8223
5_15
1062
6256
0756
8_79
3710
2453
5209
9020
8_n
.jp
gTB
T to
Wh
en
th
e H
ou
se w
as a
N
A13
4910
8054
2889
7532
154
Was
hin
gto
n M
on
um
en
t N
atio
nal
Mo
nu
me
nt
2147
7385
1
55FA
LSE
1715
6471
9446
5760
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/6
d3e
001a
27fe
eb
5cf0
14f3
2bd
f35b
96c/
5B20
4960
/t51
.288
5-15
/e35
/278
9318
8_20
5015
5718
5996
40_1
3840
5514
3871
2832
00_n
.jp
gFA
LSE
BfP
NA
iSB
xUn
1518
7411
04h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/6d
3e00
1a27
fee
b5c
f014
f32b
df3
5b96
c/5B
2049
60/t
51.2
885-
15/e
35/2
7893
188_
2050
1557
1859
9640
_138
4055
1438
7128
3200
_n.j
pg
#am
eri
can
mad
e #
geti
tto
geth
er
#tb
t #v
ega
nch
efi
ete
f #t
he
wh
ite
ho
use
NA
640
640
2844
7987
100
28Th
e W
hit
e H
ou
se11
9144
1824
2768
80
56FA
LSE
1715
6174
2818
8060
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
c016
60cd
89f3
98b
e87
2266
fa50
084d
b/5
B05
D5F
3/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
117.
0.84
5.84
5/28
1568
27_1
4600
5272
7454
13_3
5043
3261
3592
5473
28_n
.jp
gFA
LSE
BfP
GP
YRn
nYM15
1873
7556
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
8237
3773
63fe
c467
7284
819f
2de
3141
/5B
2060
E3/t
51.2
885-
15/e
35/2
8156
827_
1460
0527
2745
413_
3504
3326
1359
2547
328_
n.j
pg
Fuck
a g
lass
of
win
e, l
et'
s p
un
ish
a b
ott
le b
aby!
<f0
><U
+009
F><U
+008
D><
U+0
0BE>
NA
845
1080
100
2229
0617
711
100
Cap
ital
On
e A
ren
a37
2247
132
57FA
LSE
1715
5873
2018
1440
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/a
fa92
079c
0e62
a4a1
96c8
425b
021d
9ae
/5B
1398
0D/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/2
7891
237_
9192
8891
4887
972_
2079
6132
9059
3386
496_
n.j
pg
FALS
EB
fO_Z
QA
lEA
n15
1873
3967
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
f880
9ea0
74c7
fbe
3239
016a
5b35
5230
/5B
1967
6C/t
51.2
885-
15/e
35/2
7891
237_
9192
8891
4887
972_
2079
6132
9059
3386
496_
n.j
pg
Fin
d m
e o
n T
ind
er
NA
1080
1080
299
1122
7974
3733
299
Smit
hso
nia
n N
atio
nal
Mu
seu
m o
f A
fric
an A
me
rica
n H
isto
ry a
nd
Cu
ltu
re26
2515
071
58FA
LSE
1715
5820
9736
7630
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
5d72
5de
5ec2
7b6e
d14
ed
b4f
151f
ac6b
/5B
25E8
0E/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
181.
0.71
7.71
7/27
8795
28_5
1802
5298
5678
39_8
3614
1122
3404
6750
72_n
.jp
gFA
LSE
BfO
-NP
4ngn
m15
1873
3344
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
201a
231d
a9b
e8b
4dd
637f
a302
66a7
be
/5B
0E09
56/t
51.2
885-
15/e
35/2
7879
528_
5180
2529
8567
839_
8361
4112
2340
4675
072_
n.j
pg
SPEC
IAL
INSP
IRA
TIO
N E
DIT
ION
: N
A71
710
808
1943
9205
50
8R
en
wic
k G
alle
ry37
3555
59FA
LSE
1715
4402
2716
5480
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/6
1559
965f
9e98
2279
6a69
94cd
ed
3eb
08/5
B25
819C
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/275
7991
9_15
7949
3385
4683
64_5
7591
0838
0052
4226
56_n
.jp
gFA
LSE
BfO
d8w
9nU
tM1518
7164
32h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/98e
0c69
c482
a1a5
ea3
9a64
6a61
1d2f
03/5
B20
32B
6/t5
1.28
85-1
5/e
35/2
7579
919_
1579
4933
8546
8364
_575
9108
3800
5242
2656
_n.j
pg
cele
ry <
f0><
U+0
09F>
<U+0
091>
<U+0
0A8>
<f0>
<U+0
09F>
<U+0
08F>
<U+0
0BC
><U
+200
D><
f0><
U+0
09F>
<U+0
08D
><U
+00B
3>’s
sal
ad w
ith
ho
me
mad
e D
ijo
n v
inai
gre
tte
#tr
ynab
eh
eal
thy
#ne
win
ven
tio
n #
che
fbo
yarc
hri
s <f
0><U
+009
F><U
+009
4><U
+00A
A><
f0><
U+0
09F>
<U+0
08D
><U
+00B
4>#t
asty
#ta
keth
atka
le #
stil
lhad
#cu
red
me
ats
NA
750
750
2518
8238
192
225
The
Mal
l (W
ash
ingt
on
DC
)23
6471
522
60FA
LSE
1715
4400
8152
2620
000
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
0fd
2b96
f407
bc3
ace
9f21
82d
b59
74e
c/5B
1CD
5AB
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/275
7960
7_18
7440
4351
8045
3_41
8421
6680
8948
2444
8_n
.jp
gFA
LSE
BfO
d6p
Un
4FI
1518
7164
15h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/526
7cf9
c796
73e
ac16
05b
27e
b59
6084
b/5
B23
E4C
A/t
51.2
885-
15/e
35/2
7579
607_
1874
4043
5180
453_
4184
2166
8089
4824
448_
n.j
pg
Do
so
me
thin
g to
day
th
at y
ou
r N
A10
8010
8021
3216
0014
021
Smit
hso
nia
ns
Nat
ion
al M
use
um
of
Nat
ura
l His
tory
2146
3808
6
60
Appendix E
me
dia
_pre
vie
wth
um
bn
ail_
src
is_v
ide
oco
de
dat
ed
isp
lay_
src
cap
tio
nvi
de
o_v
iew
s
dim
en
sio
ns
.he
igh
t
dim
en
sio
ns
.wid
th
ed
ge_m
ed
i
a_p
revi
ew
_
like
.co
un
to
wn
er.
id
com
me
nt
s.co
un
t
like
s.co
un
t
gati
ng_
in
fo.d
esc
ri
pti
on
gati
ng_
in
fo.g
atin
g
_typ
e
gati
ng_
in
fo.t
itle
AC
ofy
oLl
kxz0
/L8q
3Ib
6KQ
Ycb
T6jp
/jW
LHD
U4R
V4J
wT6
1DSZ
abR
vSrC
43A
g57V
Sa2U
rlar
KN
verA
uP
KX
c/3R
S1W
wb
7kD
WR
K7s
ZBq
ob
atO
O7+
1Dch
K7e
Nvp
6fn
Tdj+
tVzs
XIZ
yP8A
p60
kw83
GTj
Ht/
Oq
4kN
Th+9
MQ
oTP
Un
g8f5
zUFx
MC
CgJ
b19
KlU
/wB
7FU
3UK
SPe
mIm
tp/J
46ZP
Pp
V/7
afV
fzrF
ALH
HrS
Hg4
PU
UrJ
ju0f
//Z
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
d7e
f33d
ec0
d40
52e
5156
eab
32d
bf9
1d/5
B0E
34A
1/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
138.
0.80
4.80
4/27
8913
57_1
8966
0747
0652
161_
4240
3877
3636
5694
976_
n.j
pg
FALS
EB
fL9J
0Jn
18a
1518
6321
29h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/8f4
eac
e78
c057
1693
6319
6ef1
997f
1c6/
5B23
CA
1A/t
51.2
885-
15/e
35/2
7891
357_
1896
6074
7065
2161
_424
0387
7363
6569
4976
_n.j
pg
A m
em
be
r o
f th
e 8
3rd
N
A80
410
8022
983
1.97
E+08
6022
983
NA
NA
NA
AC
ob
zIm
xVtX
zVA
OA
OK
sxG
gC+G
9Kvx
Sjb
yeay
TIE6
9ew
qxB
kfM
3f/A
D+l
S5Jb
jsW
J5c9
Kre
c3rQ
7A9K
h4p
3EYq
mrs
R5r
OTr
V+C
hgj
RQ
ZrSV
VC
cVn
x1b
z8u
Kxv
rqaF
STq
TVYt
Vh
6oE8
1aZL
P/Z
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
d7b
80b
4acc
801d
c902
d25
b87
ce75
33d
/5B
03B
E18/
t51.
2885
-15/
e35
/c15
8.0.
571.
571/
2758
0435
_204
1552
8327
3171
1_39
8435
6171
6863
4675
2_n
.jp
gFA
LSE
BfJ
q3d
Zn7m
G15
1855
5432
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
8543
e43
9d03
5515
aa0c
1116
b3c
2299
f/5B
076E
E1/t
51.2
885-
15/e
35/2
7580
435_
2041
5528
3273
1711
_398
4356
1716
8634
6752
_n.j
pg
An
F-2
2 R
apto
r fi
ghte
r je
t N
A69
410
8022
445
1.97
E+08
6522
445
NA
NA
NA
AC
oX
ywW
EQK
5zk9
6bu
ldSv
JB/o
afC
2YyP
Qn
9R/j
VyV
GV
Vzj
BA
I6Zw
ce9I
Cq
jov3
txP
tx/W
oG
kye
MEe
/X86
0Nw
Xjp
/n8a
ryh
WB
PR
vWn
YVxW
mEi
kY6D
Of8
AP
aq2w
/5FO
RR
t9W
BG
Pf1
H5d
qlK
xHn
JHtm
kMh
iyjE
Hp
0/w
q8L
gqC
mP
Qb
h1x
+Pe
iigY
wsr
cnI/
nV
eR
0PC
jOK
KK
LhYi
Hzj
3Hao
9z9i
fzo
oo
A/9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/532
056e
b19
cf53
a004
de
d0f
8141
d19
ab/5
B06
B31
E/t5
1.28
85-1
5/e
35/c
236.
0.60
7.60
7/27
8782
66_1
8979
7354
9485
33_5
0468
7817
9604
5619
20_n
.jp
gFA
LSE
BfG
-Gd
5HP
Gz
1518
4648
53h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/11e
9687
dca
2351
5a8e
43e
0149
8170
ee
3/5B
1322
62/t
51.2
885-
15/e
35/2
7878
266_
1897
9735
4948
533_
5046
8781
7960
4561
920_
n.j
pg
An
Ind
ian
a A
ir N
atio
nal
N
A60
710
8022
609
1.97
E+08
6022
609
NA
NA
NA
AC
ocy
FgZh
8u1+
f4SM
/kcH
FRG
KTO
znk4
x70u
wjn
jt/w
Dq
p8e
6Jtw
AJH
IzyB
9R3/
Gn
dB
YvR
6cW
QM
eT6
8j8P
Tjp
UM
un
Mh
IPB
+v/A
Nar
A1S
YIF2
JgfX
+We
uaj
bU
ZW5K
gn1y
f/r1
on
C+q
09W
Sub
qZZ
Urw
eD
SbT6
GrU
0rSs
JMb
DjH
BP
OO
/PT8
OK
h3P
/kV
k7X
02K
NC
KYE
4IP
Y+2a
bJK
I/u
gfd
Hb
PN
XP
Nd
hyS
fxP
09aj
MYb
rn/8
AV
UW
bd
x3M
wzt
1wO
evA
5/D
1pV
uG
PA
Aye
On
+eu
a0jF
7n9P
f29z
Ukc
AP
O45
HI+
7/A
PE+
9VYL
lESd
dxA
bvj
np
/So
fMH
t/n
8a0X
Ji5U
8j1C
/4V
XN
1J6j
/vlP
8A4m
lr5B
c//9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/5f9
c7b
b9a
781a
5428
7598
aa1a
2f2c
8c3/
5B19
30B
C/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
166.
0.66
7.66
7/27
5821
28_1
6501
9894
1408
76_1
4535
8459
1309
7666
56_n
.jp
gFA
LSE
Be
_ETm
vnB
4Q15
1819
9671
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
230c
9a55
740c
f81a
f2d
0175
ef9
4fb
5c/5
B1F
2703
/t51
.288
5-15
/e35
/275
8212
8_16
5019
8941
4087
6_14
5358
4591
3097
6665
6_n
.jp
g#I
CYM
I: U
.S. A
ir F
orc
e
NA
720
1080
1737
81.
97E+
0852
1737
8N
AN
AN
A
AC
oW
vwIM
Ve
VB
XP
3N+Y
TsTG
R3P
Pp
24ye
fXA
q9p
940w
w2M
jB46
EHP
bn
BB
Bzz
j0o
A1P
lBx3
pJA
FBY9
AM
mo
nm
Ctt
PfJ
B7Y
Az6
dh
71Tv
rlQ
Aw
+ZcA
gdA
Sem
f/A
K/S
gCC
0mN
wrE
9jkd
Oh
7fh
696c
YxW
ZHe
lnw
cDn
AZB
tIP
65B
6c/l
Wzg
UA
Ycsq
k+Yu
d2M
c4x2
579u
MY9
6W2l
SFy+
WA
Gd
qrx
n03
HP
Qe
lFFA
GlB
2gO
Cx7
niq
Fzcr
Ou
Mtu
Un
rjaR
2BG
eo
9cU
UU
AU
otu
7LZw
Oe
Of8
Pz5
rQN
+vo
f0/w
AaK
KA
P/9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/ad
d6e
5b8a
7c68
022a
1841
0c2e
464b
d29
/5B
0D57
70/t
51.2
885-
15/e
35/c
182.
0.41
6.41
6/27
8793
28_2
1114
7396
2414
733_
6343
2384
3641
2981
248_
n.j
pg
FALS
EB
e8k
pU
5HU
Yw15
1811
5963
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
048b
60b
0bd
74d
0ca4
a235
6b92
c7ce
31/5
B17
B1F
A/t
51.2
885-
15/e
35/2
7879
328_
2111
4739
6241
4733
_634
3238
4364
1298
1248
_n.j
pg
Sen
ior
Air
man
Ad
am
NA
576
1080
2009
01.
97E+
0870
2009
0N
AN
AN
A
AC
ob
zj8g
LAn
jt1/
z+d
VR
NID
nP
+H5V
NLJ
xsH
frV
fdzx
TYFs
XIw
GI+
bo
V9f
en
+dlf
k5d
ufp
9f0A
He
oH
jC4D
de
uO
2P88
c/h
6UsJ
H3Q
Src9
em
Pb
0469
MD
pya
ALK
xEA
An
Ptn
v3ye
/6D
2o2D
0H5U
8bgO
Rke
3H6H
+QN
QFx
nv/
3yaY
ihn
1qaJ
sHO
A3s
f/r1
X70
oP
NSM
syK
eW
bls
gE/4
f0H
4mo
I8d
/X8R
7in
qSe
vNIo
Azj
tn+l
AG
hC
zbP
m5x
wC
O/1
9/8A
PW
pM
4psX
3AfU
UV
Qj/
9k=
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
a234
46ca
8a59
8103
e35
b84
86b
7f69
66/5
B1F
32D
D/t
51.2
885-
15/e
35/c
138.
0.50
4.50
4/27
5754
16_1
6115
8280
5601
325_
5179
2782
9462
4763
904_
n.j
pg
FALS
EB
e56
xZxn
Jnt
1518
0269
00h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/eb
474b
f282
9c6d
221a
2625
9b86
6381
7f/5
B11
4892
/t51
.288
5-15
/e35
/275
7541
6_16
1158
2805
6013
25_5
1792
7829
4624
7639
04_n
.jp
gA
irm
an 1
st C
lass
Co
lin
N
A69
710
8011
094
1.97
E+08
2511
094
NA
NA
NA
AC
ocF
ZZA
Cv8
A9a
rMM
vkyA
5yD
wR
9ayP
JYZa
Fsd
9p/p
UsU
gEZc
sXlG
Cq
4x7e
hzj
r6U
hm
0L1h
k4zg
Dg/
lnvS
tfsD
938i
MH
9KxE
vZA
flIy
eO
grV
tX5P
nFV
6bR
xjvn
nn
nn
kcV
RI2
S6m
Yhl+
6f4R
/ME9
/Ufl
R5k
3qal
ku4l
O0Y
/Dkf
pV
f7e
vofy
NIZ
Qjc
Dp
VlA
FHyk
qC
cnH
H4f
SsZH
OM
1ZEr
Y+h
pA
LdQ
BT8
h68
n1/
Om
hm
Ch
Tkkd
aFO
7k9e
f04A
+lR
K5L
H2z
igY9
X69
B35
oEy
+sn
/fIq
Pfg
Dgc
5PT0
NW
gq+g
oQ
H/9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/59e
c31c
c12a
c120
dcf
7af9
7051
4686
0a/5
B1E
0534
/t51
.288
5-15
/e35
/c12
9.0.
521.
521/
2687
2232
_964
4494
4037
7403
_356
4498
5856
4915
200_
n.j
pg
FALS
EB
e3U
0YsH
xOv
1517
9398
93h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/761
530b
d5f
eb
86a6
5ee
22e
5858
9e94
15/5
B16
AB
80/t
51.2
885-
15/e
35/2
6872
232_
9644
4944
0377
403_
3564
4985
8564
9152
00_n
.jp
gA
irm
an 1
st C
lass
Kay
lon
N
A72
110
8023
499
1.97
E+08
107
2349
9N
AN
AN
A
AC
oc3
aWko
piA
nFV
mn
2MSx
+XH
Tvn
P+H
vUzn
ovd
v5D
r/h
+NU
bsF
iEH
BY4
oA
up
Mrq
GG
Rn
1FO
3Ch
VC
AK
OgG
Pyp
aAIp
JhEu
5gcf
41C
8219
pB
zgn
8B/n
ipzU
ZhB
bfz
uIx
+FA
yl9v
GS2
OcY
GfT
qfz
P8h
VcX
Ekr7
l5K
85x0
A/S
rTW
Efq
35j/
AA
qaG
3W34
Qn
n15
/pQ
BN
HO
JV3q
rYJw
On
59e
lS7v
Y/p
/jU
eT6
/yp
uTQ
B//
2Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
c4ce
dd
449d
4551
3c8d
8ce
5c20
7415
0f/5
B05
DB
F8/t
51.2
885-
15/e
35/c
129.
0.52
1.52
1/27
5805
99_4
0923
9569
4998
25_4
5065
8466
4000
9543
68_n
.jp
gFA
LSE
Be
0wcC
Qn
AE7
1517
8537
10h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/8f3
aef2
107c
703b
558a
9cc3
80ae
7934
e/5
B0D
F28B
/t51
.288
5-15
/e35
/275
8059
9_40
9239
5694
9982
5_45
0658
4664
0009
5436
8_n
.jp
gA
irm
en
ass
ign
ed
to
th
e
NA
721
1080
1902
31.
97E+
0838
1902
3N
AN
AN
A
AC
oq
sgc1
Pu
YHjk
+me
BU
J+X
n0p
qS7
VD
N1P
XkD
n05
IrN
lRV
y4H
k/iw
ao3s
Ku
vmK
MM
Ou
O4/
+tV
iScI
oY9
D/A
J7Zp
UP
me
4Ip
IuSX
Qw
RTa
fSV
qZH
QM
Oc1
AY0
Jww
Bx0
q0y
5Gaq
l8sF
9qh
lx3H
KV
Hyn
HH
0qX
d1+
hxU
KQ
qO
cVW
+2FJ
WB
GQ
OB
j/P
rUm
ktjP
Ap
Kll
KsS
ydD
zUG
6tD
An
kvZG
BO
/AP
QA
j/P
6+9H
2tQ
A+R
vGB
1HIB
6+tZ
fmN
gcn
r6n
0Ao
3t6n
86LF
XO
uR
wU
3jo
RkV
g8k5
PU
5J/O
taD
/j0H
/XP
+lZT
8Yxx
x/Q
VC
NH
0J16
c9P
XH
+f0x
Vcw
k8jp
UzE
8/W
lxV
EM//
2Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
c238
7e2d
d56
985f
3756
dc8
2647
3242
3/5A
8CA
812/
t51.
2885
-15/
e15
/268
6648
8_25
1041
6421
0100
0_49
8247
9263
8746
7878
4_n
.jp
gTR
UE
Be
wiR
jxn
116
1517
7121
21h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/2c2
387e
2dd
5698
5f37
56d
c826
4732
423/
5A8C
A81
2/t5
1.28
85-1
5/e
15/2
6866
488_
2510
4164
2101
000_
4982
4792
6387
4678
784_
n.j
pg
#Re
po
st
7638
264
064
092
541.
97E+
0887
9254
NA
NA
NA
AC
oh
fITF
gn5s
ngH
jJw
ep
9O5O
OlQ
pLG
EaX
O45
we
2WP
17D
t6D
2FV
44Q
JMX
Eg2q
OzE
nP
oO
CP
rU5e
2hU
rCA
5PH
IPH
uc8
H6Y
+tau
Texm
lbcV
Jxb
5mB
3u5G
4j0P
YHH
b25
PW
mH
VJJ
M7f
k/n
+fr6
VJJ
Is3y
Rg7
VC
Nz6
g4O
aqi2
wM
Y6jn
64P
9Tn
8Kz3
1LJB
qLI
wyT
9c8e
34f1
96T+
0nH
8T/m
f8aa
0R8x
WA
4UYP
4Cq
L27l
jgH
GTj
6UW
C5d
S7aL
jO5R
wW
x3P
PP
NSQ
3hlk
+bG
OgP
QD
HfH
vUU
du
20O
/Icr
gdTw
OM
/XN
RH
Kyf
MN
xVu
QC
Fzxy
M+n
4c1i
1v+Z
on
axo
z3W
9HQ
ZQq
6ncD
1BG
OM
Y/z7
1UO
8HG
9/zP
8AjT
QV
O5R
8oyp
Azu
wM
9Mip
toyc
MO
e2D
Wq
1Iu
RZf
++/6
/41s
wlj
Gu
Sfu
jv7V
lYJ9
D/n
3Fb
MW
Ai/
Qfy
pM
ZXt/
vj/A
HB
/Ksa
b/W
H/A
Hz/
AC
oo
qW
Ihtf
4vw
/8A
t9z9
TRR
WiA
lHSt
teg+
lFFS
wR
//2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/21a
825e
98e
f97b
f87d
af6a
e6b
04b
be
b3/
5B21
24D
5/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
94.0
.756
.756
/275
7839
3_17
8842
0838
1202
90_3
5298
2383
2904
8268
80_n
.jp
gFA
LSE
Be
vw1f
gHt8
y15
1768
6146
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/0
f75c
aeb
8377
8568
b72
330d
04af
d1f
5f/5
B03
C27
4/t5
1.28
85-1
5/e
35/2
7578
393_
1788
4208
3812
0290
_352
9823
8329
0482
6880
_n.j
pg
#Re
po
st
NA
864
1080
1760
31.
97E+
0839
1760
3N
AN
AN
A
AC
oc5
wC
nb
STgd
aBU
/nED
aMA
Y9P
60ih
JIgm
Bn
OR
UU
iFT7
U4j
1pd
5UcU
ySIi
m1K
1Mq
UU
xy0O
vem
Cn
UA
SFG
PYf
nTC
pzg
0uTS
Gi7
HZC
sDU
dLm
koEz
//Z
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
4bf7
075d
3ea6
4285
a0d
05c0
8514
6502
/5B
06D
64A
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c17
7.0.
726.
726/
2757
4285
_150
1981
0235
6690
_149
2403
9703
1486
2592
_n.j
pg
FALS
EB
etA
ms2
n3t
C15
1759
3751
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/f
fb33
8313
9266
6da8
c563
4128
8a38
650/
5B11
5E73
/t51
.288
5-15
/e35
/275
7428
5_15
0198
1023
5669
0_14
9240
3970
3148
6259
2_n
.jp
gA
irm
en
ass
ign
ed
to
th
e
NA
726
1080
1818
81.
97E+
0848
1818
8N
AN
AN
A
AC
oh
uW
8zM
7ByN
u4h
Oxx
/+o
dTV
hZ0
kDH
+4cZ
9cd
SPas
KA
u8g
jXA
Kn
5jzy
B1H
9M9/
WrF
sN6u
B8o
2t7Y
z0+v
T8q
m7K
sE07
XEJ
Db
WYS
DA
Xu
oyM
98/W
qiK
u0x
sMjI
x/vf
T0I6
/hU
NvI
qK
xIyC
oH
0yR
z+lK
0pTl
TgK
Px5
BP
v6U
0I0r
mX
zQp
7DG
Cu
cBs4
LHjo
OgG
c5ye
lZcj
FmJy
vJJ6
j1q
7bR
oV
DzZ
DH
JCq
3GP
U/X
0/G
mm
3tc/
xfm
P8K
GxJ
WN
BQ
x6E+
uM
8/jU
gDD
ucf
Wq
Uay
btq
kkjH
5YH
9Sc0
t6JA
mN
xcZ5
HP
T/8A
XV
89+h
PK
SySs
3+r5
2MjE
dyO
pB
9ulU
D9o
aTLH
cM5w
DgH
247d
qfC
SqsW
6gA
ZH14
571I
oU
/KP
lLY2
kk4y
OvA
9eu
Oxr
Ju7N
YpL5
Clx
J98n
cOu
wkD
k+g7
jpV
9U4H
J6C
qM
sLo
Cy/
eP
bI7
fXjG
evO
R6V
Msp
wN
xTd
3+Yd
e9L
Qe
vQsx
fcf6
f1N
U5O
lFFa
IzZU
k+6f
qn
8zW
jp/+
u/4
Cf6
UU
Vm
916m
i2fo
O1j
/j2P
8AvL
XJU
UV
Zmf/
Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/4f5
811f
6e65
676e
260c
70f4
6900
51e
8f/5
B0E
2EC
8/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
94.0
.756
.756
/268
6527
7_16
5327
7974
2691
0_58
0440
9423
0529
3107
2_n
.jp
gFA
LSE
Be
qG
dC
2n-9
o15
1749
6154
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
72b
3418
ffd
1624
d41
9ff6
4f37
bd
e83
6/5B
0712
B1/
t51.
2885
-15/
e35
/268
6527
7_16
5327
7974
2691
0_58
0440
9423
0529
3107
2_n
.jp
gD
eve
lop
ing
Inte
rnat
ion
al A
irm
en
wit
h t
he
ski
lls
and
cap
abil
itie
s re
qu
ire
d t
o c
on
du
ct s
ecu
rity
co
op
era
tio
n d
oe
sn't
hap
pe
n w
ith
ou
t te
amw
ork
an
d a
to
tal c
om
mit
me
nt
to o
ur
inte
rnat
ion
al a
llie
s an
d p
artn
ers
. Ou
r A
irm
en
are
als
o w
ork
ing
aro
un
d t
he
wo
rld
to
pro
vid
e s
ecu
rity
, hu
man
itar
ian
re
lie
f an
d e
du
cati
on
wh
ile
lear
nin
g in
valu
able
less
on
s fr
om
ou
r p
artn
ers
acr
oss
th
e g
lob
e.
#Par
tne
rs #
Inte
rnat
ion
al #
Air
me
n #
Skil
ls #
Co
op
era
tio
n #
Team
Wo
rk #
Co
mm
itm
en
t #A
llie
s #G
lob
e #
Re
lie
f #E
du
cati
on
#P
OTD
#P
ho
toO
fTh
eD
ay #
Air
Forc
e #
USA
FN
A86
410
8014
097
1.97
E+08
2714
097
NA
NA
NA
AC
ob
0glZ
V5f
GP
5Yzj
HU
+v+A
/nU
c15O
G2E
hT/
AJ7
1lT5
JOat
u5C
Vh
Tcu
SW6/
WtK
01A
kgP
yPf+
lZcW
e1P
iBe
RQ
Ou
Rz+
P8A
hU
37Fn
WB
cjI5
B6U
7ZW
PcX
k1vK
V4A
4I5z
1/T9
KlF
7cH
+Afm
tXzG
fKSz
WQ
lcP
I3A
4xjH
6k1m
Pae
YSYz
uV
CQ
cde
Dx9
fr+V
MZA
c59c
dT7
VC
kjR
PlC
VO
e31
rM0J
WtH
OSi
sBw
OQ
Ovf
nv8
AgK
0LW
2W0H
mzH
gdfb
6e/p
U/m
sIs9
wc9
B6f
SsG
WZ5
Pvk
nG
aALk
qxX
Em5X
Kg8
ne
On
oB
jtU
32R
P+e
qf5
/Gsp
TUm
aoR
//2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/d50
cfa8
5417
db
dd
d88
ba2
70d
5bb
c977
1/5B
0E56
78/t
51.2
885-
15/e
35/c
132.
0.51
5.51
5/26
8645
35_4
2379
3951
3870
17_1
2835
6792
0055
5827
20_n
.jp
gFA
LSE
Be
n77
mV
HR
0U15
1742
3528
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
dad
0a45
b1b
ab9e
2628
f46a
7440
b6e
db
/5B
0BB
E07/
t51.
2885
-15/
e35
/268
6453
5_42
3793
9513
8701
7_12
8356
7920
0555
8272
0_n
.jp
gSt
aff
Sgt.
Mic
hae
l N
A71
310
8021
184
1.97
E+08
8921
184
NA
NA
NA
AC
ocj
SKrc
KI5
2fcf
+6e
/up
7j9R
3Ap
klxD
AM
swJH
Ycn
/634
01rm
CYf
eG
V+Z
Q4G
AR
24O
ec8
Hr3
HSt
G+x
KR
eN
pV
Sa32
mp
bb
UW
eLf
IFR
Bzk
kk4z
jgd
Tzxk
mo
rnU
IVA
YZb
cTx0
OB
1PP
6etS
mO
xSe
Ko
PK
rYd
Uxn
IGe
eo
qvi
P8A
vL+Y
q7o
k50I
x7U
/yzk
A4H
0p4N
IpJy
e+c
fhW
JoSp
lTtz
xjp
/n3p
hTJ
9aaD
834V
JuO
KQ
xmB
Tacz
Gm
5oA
//2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/e85
a2c2
61f7
f804
2a4c
277b
7961
ab82
b/5
B03
E460
/t51
.288
5-15
/e35
/c12
9.0.
521.
521/
2628
2982
_178
7670
1289
0047
_888
1599
3915
0531
7888
_n.j
pg
FALS
EB
elX
trG
HID
v15
1733
7431
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
798d
7f03
9f6c
338a
68f9
026e
b50
1d98
/5B
215F
13/t
51.2
885-
15/e
35/2
6282
982_
1787
6701
2890
047_
8881
5993
9150
5317
888_
n.j
pg
Tech
Sgt
. Mat
the
w
NA
721
1080
1555
91.
97E+
0837
1555
9N
AN
AN
A
AC
oX
5mn
haE
Ga0
bW
181u
fujk
/QV
SQrl
J4ti
gnq
ece
3/1+
3t9a
hq
7cZZ
iT3q
oR
RYL
jKK
U0l
SMkj
YKe
a3Lb
ULa
ON
g27c
Rjp
/wD
Xo
op
gZs0
8bn
5c/l
/9e
qp
YGii
i4rD
DSU
UU
hn
//2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/0c1
f795
0024
a89f
f2a4
13ae
e7c
1ab
f01/
5B00
D0C
1/t5
1.28
85-1
5/e
35/c
170.
0.43
9.43
9/26
8662
99_2
4667
9619
2048
74_2
4933
3909
7850
9701
12_n
.jp
gFA
LSE
Be
iyW
wfn
Vw
Q15
1725
0736
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
7745
9429
dc6
f2fd
438d
7c03
da4
e05
15/5
B28
163A
/t51
.288
5-15
/e35
/268
6629
9_24
6679
6192
0487
4_24
9333
9097
8509
7011
2_n
.jp
gA
n A
tlas
V r
ock
et
carr
yin
g N
A60
710
8019
678
1.97
E+08
5219
678
NA
NA
NA
AC
oX
qC
1l6k
Y/EU
/yG
X7w
x7n
gVG
t6R
zmm
STp
Kcu
Cce
5pag
WV
jyM
rg/S
mtE
chcc
kZA
/wA
9Kp
7ow
eFP
51L9
px9
1cE0
wH
vC69
qj8
pva
mtc
s3W
ofN
NA
EG/H
NO
Eho
op
gJvN
OEh
+tFF
AC
b/A
Gp
u6i
igD
//Z
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
5b80
322b
0132
8fa2
a2e
8e5e
a608
4039
/5B
0609
F1/t
51.2
885-
15/e
35/c
237.
0.60
6.60
6/26
8642
78_1
5207
1498
7240
27_5
4927
3568
5175
5417
60_n
.jp
gFA
LSE
Be
a_TS
Lnw
Wm
1516
9890
87h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/947
9866
567f
985b
f621
d0c
dd
238a
c68b
/5B
18E3
27/t
51.2
885-
15/e
35/2
6864
278_
1520
7149
8724
027_
5492
7356
8517
5541
760_
n.j
pg
A s
tud
en
t w
ith
N
A60
610
8022
522
1.97
E+08
6822
522
NA
NA
NA
AC
ocx
EUEA
EkEH
GO
vXu
P6j
8qke
M7N
ygjb
yT2w
TgYB
pU
UlD
gZA
wT6
j3q
3GA
xPP
HJO
ep
47+u
Sf61
s46e
hK
etu
5nM
5Kh
cnH
cev1
qN
n4x
6Vc8
hQ
5X07
f5/S
myw
hsB
eM
VN
9NA
8mU
ghI3
dv5
n0o
qxN
A0e
Aw
6jP
FQYX
3qP
Qo
tW7P
nYv
U+/
WtK
05Eh
k6o
m0e
o9M
/0P
pW
LG5B
yO1a
1iN
4kJ7
jP86
1b0f
8ATI
j8Sv
3JLy
HaP
N3A
YAG
zu3O
CQ
fpg/
Wq
rMA
Nw
BYd
+3P
vVu
7Ym
JT67
f1A
zSO
oV
9gH
ysM
kfp
WJo
1qyi
z7+e
tRU
jfK
SB2N
Np
kn//
2Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/6
2116
6983
dd
4540
71e
7d72
4100
1c30
37/5
B08
AD
BF/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c17
9.0.
721.
721/
2636
6961
_153
3404
8872
0051
_424
9312
2875
6475
9040
_n.j
pg
FALS
EB
eYV
bl-
Hn
Fr15
1690
0027
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
71c0
98b
a4d
0b1a
a24a
307b
0f50
b92
68/5
B16
9424
/t51
.288
5-15
/e35
/263
6696
1_15
3340
4887
2005
1_42
4931
2287
5647
5904
0_n
.jp
gU
.S. A
ir F
orc
e S
taff
Sgt
. N
A72
110
8015
703
1.97
E+08
3315
703
NA
NA
NA
AC
oV
0g2R
nrS
Bw
OvW
s9Lx
T1X
b/w
AC
P+F
KZE
Y8fz
/xFQ
7Bq
Xjd
hTU
i3iN
3rK
2F+g
HH
oR
z+R
pp
Q5x
tK8n
r0x2
paF
amp
JKh
NR
7vrV
VJD
Gc4
Bp
/2s/
wC
f/w
BV
NC
Mo
U8U
UU
wG
07e
fU8U
UU
AJv
J680
uf8
5oo
oA
//9k
=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/f30
aa70
0ea3
2c6f
02a0
a3b
5bb
07b
3223
/5B
0758
63/t
51.2
885-
15/e
35/c
257.
0.56
5.56
5/26
3185
22_1
9541
8553
7986
764_
3255
8042
1075
0423
040_
n.j
pg
FALS
EB
eV
3SU
4H7L
a15
1681
7113
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/7
d32
ad62
6da8
24e
7d3b
1e12
0bb
9bb
2cb
/5B
0FB
877/
t51.
2885
-15/
e35
/263
1852
2_19
5418
5537
9867
64_3
2558
0421
0750
4230
40_n
.jp
gU
.S. A
ir F
orc
e F
-15C
Eag
le
NA
565
1080
2189
01.
97E+
0863
2189
0N
AN
AN
A
AC
ob
2tvp
zUTS
7PvK
4x3A
3D9P
8Kt0
Yp3F
Yy4b
9JG
YMC
gGM
Eg/i
D7g
9vSr
qsr
9P5H
+oq
XC
rzw
PyF
NM
ijq
R+d
AW
GlK
bsp
Gu
Ix/E
Kj+
1xe
tArG
X9v
dv4
j+FV
bm
6lO
Np
IGD
+Y/w
AK
p9B
xSkm
pu
XYk
NxL
k4P
Gz/
wD
Wfr
RB
O7M
Q3T
A/z
/jU
B4/
Wh
fvH
3zR
cLFw
vTN
9Qn
im5N
IZ//
9k=
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/d
6f9c
5563
44cc
e4f
f250
1298
c8a9
a901
/5B
22A
7A5/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c18
1.0.
718.
718/
2632
0814
_180
5266
2588
0237
_696
5410
9250
7969
1264
_n.j
pg
FALS
EB
eTR
zVjn
us9
1516
7303
52h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/d5e
601e
d61
6502
48c2
c9f9
1154
96d
d1b
/5B
0295
E8/t
51.2
885-
15/e
35/2
6320
814_
1805
2662
5880
237_
6965
4109
2507
9691
264_
n.j
pg
Me
mb
ers
of
the
146
th A
ir N
A71
810
8029
747
1.97
E+08
112
2974
7N
AN
AN
A
AC
og6
IEZx
nkV
iX2l
pP
ma3
xuB
O4D
oxH
XH
v69j
9eq
ee
La3b
5syS
Nye
pyR
yfb
0Hp
x6V
dsJ
0W3X
nn
BO
3v1P
AFA
HN
Q2v
JZj5
aDIJ
OB
yP4e
ep
/A4q
9Dcq
uV
Vh
Io/h
Y8kf
7LED
p6c
im3z
b5O
fl3E
E+n
cfm
PW
qjB
AM
rnH
Qgn
Ofp
QM
1GSN
x5gO
0dw
3BH
4f4d
aBrq
KM
CN
iBxn
Iqjd
ttX
A64
61kY
xxzQ
I1P
tW8b
QFV
fRQ
B+f
c/ia
kEu
3km
shTV
hSc
YAN
MD
Q84
PlM
ZJX
qe
gHf3
9PX
pV
NFE
f3/v
H9A
O/5
1NE7
r8u
Mo
Rzk
DIP
qD
1/C
mFZ
Ou
OR
SAb
ck7t
jkc4
II6f
5/G
qm
ypn
iZj2
FJ5U
g/io
A//
9k=
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
7620
9ed
d2a
d75
9c29
535a
02cf
940d
67/5
B1D
5F0B
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c11
7.0.
846.
846/
2622
7823
_195
0940
2652
2399
6_58
2411
5476
0363
2128
0_n
.jp
gFA
LSE
Be
Qu
T-8H
fpC
1516
6446
36h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/533
db
b35
fbd
4980
c330
27ac
7d46
b8e
34/5
B1E
1832
/t51
.288
5-15
/e35
/262
2782
3_19
5094
0265
2239
96_5
8241
1547
6036
3212
80_n
.jp
gC
apt.
Zo
e K
otn
ik, 5
5th
N
A84
610
8041
541
1.97
E+08
286
4154
1N
AN
AN
A
AC
ob
q2V
k043
Lke
6g/l
uO
B+R
zVm
bSc
8pw
e+T
1/H
PB
reh
RY0
VE4
VR
gf4/
jQw
J9P
rQB
yp0p
wM
55P
Ucc
c+vf
iqxs
HC
gnA
ycc/
XH
Ue
tdP
Icd
Tt5w
OO
Tmss
uo
Py5
bB
I6+4
H5c
1SsT
dm
E8b
RN
tcYP
8An
oab
8vvW
9JG
JFIk
GF9
xzn
275H
f8A
Wsp
rN8n
BB
HY+
1Nxf
TUFJ
dTa
sLtk
QA
yR7e
ys3z
D6Z
HT2
P51
r+aj
8hs8
diM
fpx+
NcI
KcD
ioK
Oru
MK
Gb
jPO
AO
p9M
k/0q
iW8l
Au
Qu
3HT1
HfP
rWM
ZXxj
cce
mai
LE9a
pO
xNrm
lLd
jnB
LEn
qe
n/A
Nc1
U+0
Sf3j
+dV
s0U
m2x
pJH
/2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/1f3
6f76
0f6f
1dd
a8d
ea3
3059
5772
11c1
/5B
033A
EB/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
181.
0.71
8.71
8/26
1855
73_1
4154
6253
1892
41_2
5539
3766
3308
5952
0_n
.jp
gFA
LSE
Be
I6h
nY_
C15
1638
2601
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
df8
1607
67d
31cf
478c
db
1091
3a7a
171/
5B14
7C00
/t51
.288
5-15
/e35
/261
8557
3_14
1546
2531
8924
1_25
5393
7663
3085
9520
_n.j
pg
A U
.S. A
ir F
orc
e F
-15
NA
718
1080
2432
71.
97E+
0867
2432
7N
AN
AN
A
AC
ob
3AQ
wyD
keo
5pcV
hJa
FpX
W2L
Re
WQ
OSc
Entx
0x75
96P
t1zA
CX
AlV
Tgkc
8+h
I4z9
RV
3M7C
tqrZ
ICjj
O32
xnr/
kVZi
L3d
rtB
AYj
bk9
OvP
6Vgy
kM5Y
AA
Hn
GTj
n3+
tbN
ldR
gMw
+g5b
JPp
SXm
Mvx
26xI
sa5A
Ug5
HU
kdz6
57/l
U1K
On
r70l
WSS
Vl3
dss
Fsyx
5AJ3
Hvk
57+1
adB
AIw
eQ
aQzk
ob
Z7g4
QH
qM
nsP
c/zr
SgsE
mZ9
6GM
KQ
q4Y
56ck
5yO
eD
wM
Vsr
Gq
H5R
gYH
A6d
+3Sn
UrB
cwre
x3s+
xpIw
jYB
z19e
Bir
f2O
b/n
u35
f/X
rRN
MzT
sB//
2Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
e93
0cfb
7545
0798
0d56
45b
5c1a
3b00
f/5B
0C63
D3/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c18
1.0.
718.
718/
2627
3992
_153
3488
9544
5528
_357
3194
8722
2133
5552
_n.j
pg
FALS
EB
eG
eW
ntH
ymU
1516
3007
25h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/699
2c68
a252
3de
8aaf
b2d
216a
3964
2c9/
5B1D
409E
/t51
.288
5-15
/e35
/262
7399
2_15
3348
8954
4552
8_35
7319
4872
2213
3555
2_n
.jp
gP
arat
roo
pe
rs a
ssig
ne
d t
o
NA
718
1080
1994
71.
97E+
0866
1994
7N
AN
AN
A
AC
oV
2Hm
SL7x
4HX
24yM
j36D
3Iq
VJU
Zd2Q
MD
nn
OP
bP
TNc6
LGW
Z9r5
EZyA
2/d
j0P
Jz16
ip57
Vo
IQ2S
23A
Pzc
Ae
y47H
jrm
qEW
3u5G
JMfC
np
x/K
rsN
wkq
7iQ
pH
3vQ
f/W
rlH
uG
4x2q
1bA
XD
EMSo
6ZB
xyfw
Ofp
T0Fq
bo
uEZ
tvQ
45H
U5y
RgA
de
mc9
MYP
ep
tprn
Z7Jo
2zES
xJO
TnaN
vGA
QP
Xn
P+N
PEK
kch
c9/m
k/xp
AU
hIT
x/h
Ury
OyF
Cfl
PG
MU
UV
JZSa
Lnr6
Dp
9ff2
qe
NTH
9045
z/n
mii
kBM
80h
6sah
3N60
UU
wP
/2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/1cd
c63f
2af2
34e
ee
438c
6075
672f
74d
7/5B
2391
42/t
51.2
885-
15/e
35/c
257.
0.56
5.56
5/26
3349
82_1
8197
4973
8045
330_
5625
8171
8816
7319
552_
n.j
pg
FALS
EB
eD
6eR
Jn3J
A15
1621
4804
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
75d
dc5
d21
4bfe
13cc
28c5
527d
443e
4f/5
B28
4C56
/t51
.288
5-15
/e35
/263
3498
2_18
1974
9738
0453
30_5
6258
1718
8167
3195
52_n
.jp
gA
153
rd A
irli
ft W
ing
NA
565
1080
2097
91.
97E+
0885
2097
9N
AN
AN
A
AC
oX
xY3O
PcV
biX
zRn
AJ+
uK
pxo
rHb
kir9
u8d
vkSA
uu
Plx
1yT9
Rm
oe
uh
a01E
uIS
qb
mX
Zzgc
57H
0qkF
DD
8efp
W4Z
IJV
+ZW
2jn
v16d
nq
ncR
wM
u2N
Sr5H
zZ4x
6YLH
mtY
6KzM
5avQ
z3jQ
HjO
PrT
cJ6G
pn
tmjy
OD
mq
/lN
SAR
Gw
c1aE
oC
hiM
kHii
ikM
RX
LKN
p24
AX
j05P
+fan
Bto
65xR
RQ
9hd
SJ7p
m61
F5p
oo
oG
f/9k
=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/3c4
782b
8a37
be
1bb
2541
305b
b36
0d55
6/5B
26F6
4D/t
51.2
885-
15/e
35/c
147.
0.38
5.38
5/26
0731
53_1
4795
0699
2437
34_3
9355
2319
2394
7438
08_n
.jp
gFA
LSE
Be
AyL
86n
6Ad
1516
1097
96h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/9d
4fb
b97
8ecc
d1f
7c79
ed
3dab
5030
3ea/
5B0A
D90
4/t5
1.28
85-1
5/e
35/2
6073
153_
1479
5069
9243
734_
3935
5231
9239
4743
808_
n.j
pg
YEA
R IN
PH
OTO
S IS
N
A42
475
010
355
1.97
E+08
2510
355
NA
NA
NA
AC
ob
jHu
Dj1
FT4C
e/+
e3r
Uq
r8m
Bg5
Pe
nFQ
OA
Ku
2rI6
FK6Y
rH7n
pW
MP
lyT+
H0q
9qEm
W2j
t/w
Dq
/wA
fyq
Dy1
dcF
sOB
jGO
uO
gz69
/pn
vUsp
FM/N
/Wn
7DU
89sY
Qu
4Fe
Bn
PX
Pcj
2xTT
IoP
BO
O2Q
KSQ
HX
C3I
9x7U
kkO
1Sx4
xzn
6c/r
0q/i
qO
pEi
3bH
t/M
UD
OSl
k3Sb
jzzn
8v8m
pC
SwyM
KFA
Ixw
x9Tn
uf/
1iq
TH+Q
oyc
U0J
kssh
Y4JJ
+pyf
zNQ
ZoN
JQB
/9k=
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
36d
1442
c059
3b9a
09ab
c3c6
bd
58ce
d5/
5B07
3FB
8/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
181.
0.71
8.71
8/26
1518
58_1
9523
5251
8424
829_
8116
8030
3478
2105
600_
n.j
pg
FALS
EB
d2-
Z9h
Hrl
N15
1578
0658
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
22cd
96c5
ce38
7dd
0cc7
ad11
f9d
7eb
e6/
5B18
A75
B/t
51.2
885-
15/e
35/2
6151
858_
1952
3525
1842
4829
_811
6803
0347
8210
5600
_n.j
pg
Cap
t. M
ich
ael S
war
tz,
NA
718
1080
1638
31.
97E+
0861
1638
3N
AN
AN
A
AC
oW
pD
nrT
g3JH
pj+
VM
J2jP
YU2P
JYvy
B0w
fp1o
AtA
07N
Rkj
GO
+f0o
Bo
AlZ
Qw
wR
kGq
RsY
/Vv0
qck
qcj
kf5/
rT9w
oA
rYzw
f84p
y8fj
zRR
SGK
eaA
SaK
KA
F3U
Zoo
oA
//Z
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
ed
3821
9955
f77f
a2ae
4c80
9d89
6883
b/5
B13
009A
/t51
.288
5-15
/e35
/c25
0.0.
579.
579/
2606
6654
_537
3303
3664
2122
_681
7855
0258
1810
7904
_n.j
pg
FALS
EB
d0Y
LR2n
kX4
1515
6935
06h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/39a
28b
d03
1737
015e
fb77
9429
d01
3ed
1/5B
080A
D7/
t51.
2885
-15/
e35
/260
6665
4_53
7330
3366
4212
2_68
1785
5025
8181
0790
4_n
.jp
gTw
o A
-10
Thu
nd
erb
olt
IIs
NA
579
1080
2489
21.
97E+
0895
2489
2N
AN
AN
A
AC
ob
xlX
FSgU
0VIp
wcj
tzSA
aBSE
Z79K
sZh
k++P
LJH
8HQ
n2U
g/iM
j8K
aIc5
54A
6j/A
4/K
gCsw
qLF
S7s5
U9R
19P
/rfS
m4F
AD
xT6Y
p6V
Ioo
AiL
Op
yv5f
560q
Tktk
jdj+
H/P
p1p
wN
QgZ
J9sm
mB
KSO
Qfv
9TjG
Pp
/njt
Ud
NLn
cR2y
aZm
gD/2
Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/0
c21d
8a13
bc7
778f
1079
05e
2002
dad
de
/5B
09E8
30/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
171.
0.68
2.68
2/26
3354
23_7
6006
6850
8487
22_8
2291
3388
0024
6292
48_n
.jp
gFA
LSE
Bd
x1La
5nv1
r15
1560
8048
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
970b
6320
dd
641e
0519
1dce
696a
ed
202/
5B01
3C47
/t51
.288
5-15
/e35
/263
3542
3_76
0066
8508
4872
2_82
2913
3880
0246
2924
8_n
.jp
gSe
nio
r A
irm
an A
very
N
A71
910
8030
813
1.97
E+08
175
3081
3N
AN
AN
A
AC
oX
D0p
A8c
fVT7
7gR
/StU
Qj1
NO
MYH
8R/P
itH
O+2
hko
W3M
1Lq
3I3Y
wP
pn
+VN
Op
QD
hA
zfgA
P1q
y1p
asf4
Sfo
B/L
FQe
TArY
8ks3
bB
U/j
14rO
5pYn
t5R
cLkE
Ejrg
HA
9OT1
/Cp
vKN
OA
XA
zkH
0z09
qM
J7/n
QB
SN2S
OO
v1N
PFw
CP
mO
fXii
itb
Izu
VJG
hb
+H9P
8DU
cLxx
52jb
74o
op
2Qrs
e10
w6M
CP
cEU
37a/
p+v
8A9a
iilZ
Du
f//Z
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
bb
9f76
46af
45cd
a8b
aacc
5382
00e
f05/
5B11
4797
/t51
.288
5-15
/e35
/c22
3.0.
577.
577/
2618
6025
_167
7093
9289
9569
0_58
8053
9848
6783
0988
8_n
.jp
gFA
LSE
Bd
vOB
ByH
aSb
1515
5204
07h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/42a
b5d
95e
a0d
eb
5e25
c40d
e7d
a3ad
098/
5B14
D9B
4/t5
1.28
85-1
5/e
35/2
6186
025_
1677
0939
2899
5690
_588
0539
8486
7830
9888
_n.j
pg
Hil
l Air
Fo
rce
Bas
e F
-35A
N
A60
810
8027
976
1.97
E+08
120
2797
6N
AN
AN
A
AC
oc5
7FIf
Qip
Qm
etN
cAd
Pxq
mh
C7a
Xb
U0a
7xm
pfL
rRIm
5TK
0m2r
Zjp
myi
wX
BFz
UzW
wkA
OcH
1rO
DG
pV
md
ejE
VH
N0H
YnW
KSL
JU9D
yp7j
1Hr+
HN
WY7
lZO
FUlg
CSO
Og6
47n
8s1R
NxI
3Via
jDsj
b1J
DD
uO
tJSa
G0m
bD
RSF
tm0K
Su4E
kkY+
oA
Gfb
NR
+Q/q
v5H
/Gst
riR
urM
fxN
R7j
6n86
OZh
ZH/2
Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/a
e27
1c4c
ff61
afd
817e
fe90
0760
a997
f/5B
032D
7F/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
166.
0.66
7.66
7/26
2231
36_1
8609
8062
1261
76_4
7795
0152
7532
3064
32_n
.jp
gFA
LSE
Bd
ssgC
on
0pj
1515
4357
27h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/117
27fe
46b
e64
8040
7f40
6cd
4c2e
17d
f/5B
19D
AC
0/t5
1.28
85-1
5/e
35/2
6223
136_
1860
9806
2126
176_
4779
5015
2753
2306
432_
n.j
pg
A U
.S. A
ir F
orc
e
NA
720
1080
1862
51.
97E+
0840
1862
5N
AN
AN
A
AC
od
50A
tgD
kmla
Jl6j
BJI
x9K
ktn
Ktk
dq
tTyh
+eh
Hf3
9vw
6/h
XW
ocy
5n/A
FYyc
mp
Wto
RTW
hjR
XB
yDgH
2Jq
K4h
8kjn
O4Z
qaS
4Lq
FGd
uR
17kf
yqJ3
3jJ5
4x/w
Dq
oaW
uvo
THm
05u
7/4B
Xo
qW
THG
0Yq
Oo
sagr
be
aUsW
680l
SAYq
km1a
+gM
tW1s
G5l
YKvp
yT+l
JN5Y
+RTu
xxn
GB
+ee
aqjJ
zzU
iRh
hz6
UR
g5P
8Aro
JuxG
w2c
Ec1H
RR
Ut9
thn
//Z
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
59a7
a0a9
014b
f4d
40e
29d
1d37
39e
ff4/
5B01
3986
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c14
3.0.
714.
714/
2503
6855
_131
8547
7949
1786
1_71
8908
5056
5545
0009
6_n
.jp
gFA
LSE
Bd
lOYo
TnU
Si15
1518
5056
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
eb
fdc1
ff79
d2e
986d
9d97
35a7
e8b
8d4/
5B1B
011B
/t51
.288
5-15
/e35
/250
3685
5_13
1854
7794
9178
61_7
1890
8505
6554
5000
96_n
.jp
gA
irm
an 1
st C
lass
Jo
rde
n
NA
771
1080
1459
11.
97E+
0845
1459
1N
AN
AN
A
NA
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
4ea5
f540
4b15
7a38
4986
4bab
21d
00fd
/5B
1F13
56/t
51.2
885-
15/e
35/c
111.
0.44
4.44
4/25
0146
24_2
0417
3480
1582
32_4
3418
8371
4855
2396
80_n
.jp
gFA
LSE
Bd
iHp
U9H
zMZ
1515
0808
60h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/648
5d36
ba7
124f
ed
48c8
7bfc
ac05
2cf9
/5B
0E0A
4F/t
51.2
885-
15/e
35/2
5014
624_
2041
7348
0158
232_
4341
8837
1485
5239
680_
n.j
pg
The
go
vern
or
de
clar
ed
a s
tate
of
em
erg
en
cy in
28
of
Ge
org
ia's
co
un
tie
s in
clu
din
g M
oo
dy
AFB
aft
er
an e
xtre
me
win
ter
sto
rm b
rou
ght
sno
w, s
lee
t an
d f
ree
zin
g ra
in t
o t
he
are
a. T
his
is t
he
fir
st s
no
w t
he
bas
e h
as s
ee
n s
ince
201
0. (
U.S
. Air
Fo
rce
ph
oto
s b
y A
nd
rea
Jen
kin
s) #
sno
w #
ice
#ge
org
ia #
c130
#a1
0 #b
rrr
#brr
rt #
sou
th #
sto
rm #
sno
wfl
ake
s #s
no
wst
orm
#ai
rme
n #
po
td #
airf
orc
e #
ext
rem
ew
eat
he
r #s
lee
t #f
ree
zin
gN
A50
075
036
611
1.97
E+08
212
3661
1N
AN
AN
A
AC
oc5
vFA
FaJt
Q5y
Dj2
qB
rd06
j8u
aAC
Gye
dd
yEEg
4I6f
Sph
pkg
5cq
n1N
Qo
7xZC
krn
rjio
yxP
J5P
vQB
dG
nL0
MgJ
PYY
/wA
ad/Z
6ep
/T/C
qSH
DA
jrkV
0RA
oA
yo7h
B1p
qX
A87
cSd
o5/
H/A
PV
WcK
cKQ
zfae
3fk4
z645
qJm
i/h
KH
6rz/
Ksf
NLk
0AaP
nB
fugD
6AU
n2g
+tZ2
TSZp
iP//
Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/3e
aae
4c55
1ce
c7ce
ecc
609c
a0cf
ceb
c8/5
B02
07E6
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c16
6.0.
667.
667/
2627
1992
_352
8718
4518
6923
_892
4461
1808
4450
7136
_n.j
pg
FALS
EB
dfu
ca_H
CA
715
1500
0538
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
0b3a
8a06
6425
bb
e06
ba5
533b
3a21
497/
5B23
FD59
/t51
.288
5-15
/e35
/262
7199
2_35
2871
8451
8692
3_89
2446
1180
8445
0713
6_n
.jp
gA
U.S
. Air
Fo
rce
cre
w
NA
720
1080
1538
91.
97E+
0838
1538
9N
AN
AN
A
AC
ob
fMd
x2js
KZ5
fyk9
h+p
qW
Q/M
ailf
bH
7D9T
Wp
mU
5HV
etX
bfB
A9q
o+W
shD
+o6V
ftB
k49K
HsJ
Giq
A+w
pD
GP
Wn
49O
TTh
FJ7f
nW
ZoZL
H5j
9KrX
B3A
A9F
7ev1
q1L
y7V
TkrR
Gb
Kq
yFTn
PX
rWta
zIil
m4L
fjW
WQ
Mn
8KtH
oP
oK
GSj
TF0r
Zxn
gZp
v2sf
5NZi
HB
b/A
HT/
AC
NU
jI3r
UG
iZ/9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/742
c5d
eb
fb7f
973a
6a29
19cf
011a
374e
/5B
09F7
C2/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c18
1.0.
718.
718/
2502
4242
_468
6105
6353
4878
_288
3504
7159
7745
7664
_n.j
pg
FALS
EB
dd
OW
Rd
nV
qK
1514
9166
02h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/fe
eb
763b
d44
2309
eb
4046
4224
b0f
8a84
/5B
2734
8F/t
51.2
885-
15/e
35/2
5024
242_
4686
1056
3534
878_
2883
5047
1597
7457
664_
n.j
pg
Elin
or
Ott
o, b
ett
er
kno
wn
NA
718
1080
1930
71.
97E+
0848
1930
7N
AN
AN
A
AC
oY5
7FLi
pTH
hN
+aG
C9j
x2zW
luh
Nxo
He
n4y
o/E
/0/p
RIM
Rgj
BO
SPfp
/Kp
ihQ
D2C
/j3q
I73Y
2MW
MH
rTd
oq
75f0
I9R
R5d
dP
KY8
xREb
+hq
R/M
YEY6
+1FF
c3M
zayJ
PK
Lgg8
fdI/
Ln/8
AV
Vp
AD
EVd
eTx
17U
UV
Cb
WxW
4pco
Aq
YOB
1PX
/PvS
ec/
ov5
n/C
iir5
n3J
sux/
9k=
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/0
476b
d73
e8d
29e
7d58
3e5a
6bd
3318
226/
5B13
4D38
/t51
.288
5-15
/e35
/c16
7.0.
446.
446/
2502
1787
_108
8851
4987
9994
_604
0470
5022
0240
8960
_n.j
pg
FALS
EB
dS2
lh1n
oG
915
1456
8599
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
bb
aa95
045a
3f08
8f78
a3e
dcc
506d
e03
/5B
162F
62/t
51.2
885-
15/e
35/2
5021
787_
1088
8514
9879
994_
6040
4705
0220
2408
960_
n.j
pg
An
Air
man
ass
ign
ed
to
N
A61
710
8019
447
1.97
E+08
6119
447
NA
NA
NA
AC
ocy
ZbiS
Y5ck
+g7D
6Dp
Vm
wn
ZSy5
+Xax
/ED
j8c1
Wm
gMX
J6H
p6/
lTIX
CN
z074
Gf0
yM07
hY6
sIA
o5z
gck+
w5z
TTG
K51
5Gly
N42
n2I
49w
Ae
Kkt
2Ecg
bcc
Y6ZP
8Akf
55q
uYm
xtG
Km
eV
VV
7jd
91j+
FV/N
b++
fzN
Pn
FylG
4zvy
3ftn
JA9/
51X
Nb
M8C
JEcD
tn8a
xAaz
NC
bd
xio
91N
yaX
FAh
6uQ
eM
ipR
ayn
nA
59x/
jVcH
B9a
0AW
Izu
b86
AP
/2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/15c
aa12
72e
acfe
fd32
44fc
6030
a0e
312/
5B1E
721E
/t51
.288
5-15
/e35
/c13
0.0.
520.
520/
2503
6988
_906
8534
6613
7934
_788
3972
8361
9110
912_
n.j
pg
FALS
EB
dQ
SJd
Yn6f
315
1448
2386
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
691f
66aa
6150
69a6
ba2
687f
d4e
fcd
8b/5
B1E
7E1C
/t51
.288
5-15
/e35
/250
3698
8_90
6853
4661
3793
4_78
8397
2836
1911
0912
_n.j
pg
Staf
f Sg
t. R
ich
ard
Mu
rkin
, N
A72
010
8020
520
1.97
E+08
5520
520
NA
NA
NA
AC
oZm
vtN
OTJ
AP
coP
5r/h
+Xp
WSG
rs6y
NTs
lZTM
vDD
72O
/v9a
lrq
jto
1mm
oP
0Ri7
qN
1R8b
ttLs
NSe
ipX
29A
NM
q3b
2kly
SExx
1J6f
TvU
v8A
ZNx6
L/31
SMJ1
Ip2b
sy//
AG
r3X
awzj
acq
w56
/xK
R+V
TjU
IZQ
Y5P
l3A
jqD
wR
1yO
n4g
Vyy
9atC
rue
Wgn
QR
SlU
O9c
cEe
h6f
/AF6
uR
x20X
+vfe
xH3V
Bw
D9R
1/Sq
MlV
jS0N
3Vn
a1/8
zcG
o+S
CsC
KEH
Izkk
n3x
3+p
/Gp
hfu
Rn
fGM
9tj/
wC
Nc2
abTM
m79
F/X
qf/
Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/cd
4879
6e4c
fd2a
e99
e56
903b
0f81
570f
/5B
20B
912/
t51.
2885
-15/
s640
x640
/sh
0.08
/e35
/c21
8.0.
644.
644/
2618
6523
_192
8207
4127
1611
_741
9771
7228
6740
8896
_n.j
pg
FALS
EB
dN
sGo
NH
OC
t15
1439
5331
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/c
ac39
2949
4f74
e60
09b
bb
5c11
0c77
ae3/
5B02
A0C
7/t5
1.28
85-1
5/e
35/2
6186
523_
1928
2074
1271
611_
7419
7717
2286
7408
896_
n.j
pg
Air
me
n f
rom
Ne
llis
an
d
NA
644
1080
1623
61.
97E+
0865
1623
6N
AN
AN
A
AC
oe
nu
tQTl
FJD
D+6
M4x
654/
Dm
pLi
dW
thIC
cMV
578H
n8e
KxJ
7gN
KxX
5e27
uSP
5Z6e
wq
68im
1ES8
kYYY
/r/n
mgZ
Z+1C
S03f
N8j
KrZ
OSc
EHO
fU+/
Sp7K
6WaM
5J/d
/eZj
k85w
TWP
vZYG
Rm
HO
TtH
B3A
/4D
8ulX
LVSb
SRyA
d7Y
I6Yx
gdR
+Y/w
NM
Rb
ivYi
oM
jAN
kg4y
Bkd
OT0
yOe
uK
ugZ
5HSu
cYA
xl3y
eFO
OD
k9O
R6N
gnjq
BX
Qxz
K6K
w4B
AO
PTI
6UC
OM
l4c+
5P86
1YZB
LHlg
Nx4
Bzw
Md
wP
6Vm
Tjn
Pb
p/O
rdvj
y8H
60D
Jop
TuyA
oG
MD
dkj
P97
B7+
3QU
kty0
KA
Lj3O
Tz35
/z+F
QD
GP
8/40
2ZQ
YwR
/e5/
Ggd
yKW
dn
BB
I5P
QZH
ToP
wxV
6K5I
RR
zwB
39q
ySc/
rWgi
fKO
ew
7UM
R//
2Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/1
2a07
5cc1
50b
06c2
2a52
607d
7712
f831
/5B
2514
8F/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
150.
0.78
0.78
0/25
0125
52_1
5556
2435
0798
04_4
2823
0650
5505
5052
8_n
.jp
gFA
LSE
Bd
BK
a9H
HJA
_15
1397
5019
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/5
0432
a2b
bb
8242
06a7
99b
ecb
e34
2a0c
2/5B
0926
FB/t
51.2
885-
15/e
35/2
5012
552_
1555
6243
5079
804_
4282
3065
0550
5505
28_n
.jp
gA
irm
en
fo
ld t
he
fla
g d
uri
ng
an h
on
or
guar
d g
rad
uat
ion
ce
rem
on
y at
Egl
in A
ir F
orc
e B
ase
, Fla
., D
ec.
1, 2
017.
Ap
pro
xim
ate
ly 1
3 n
ew
Air
me
n g
rad
uat
ed
fro
m t
he
120
-plu
s-h
ou
r co
urs
e. T
he
gra
du
atio
n p
erf
orm
ance
incl
ud
es
flag
de
tail
, rif
le v
oll
ey,
pal
lbe
are
rs a
nd
bu
gle
r fo
r fr
ien
ds,
fam
ily
and
un
it c
om
man
de
rs. (
U.S
. Air
Fo
rce
ph
oto
by
Sam
ue
l Kin
g Jr
.) #
PO
TD #
Ph
oto
OfT
he
Day
#A
irFo
rce
#U
SAF
#Fla
g #H
on
orG
uar
d #
Egli
n #
Flo
rid
a #A
irm
en
#P
allB
ear
er
#Bu
gle
r #F
amil
y #H
on
or
#Fo
ld #
Fun
era
l #C
ere
mo
ny
NA
780
1080
2610
41.
97E+
0813
826
104
NA
NA
NA
AC
oc5
zFG
K2p
0ikw
RH
tOR
k7gO
D64
zj64
qtD
bR
zZIc
Jgjg
gng9
898f
hSG
Z+K
TFb
C2E
Z/5b
KO
SOh
7dD
171C
9tG
i7jJ
n2C
k8/n
29aL
gZw
Un
gc0b
TW1b
sIC
Cq
hgy
nlW
PP
IOTu
HG
Om
KlM
kTEk
xNk8
n5l
/xp
gY7z
k9O
lRg8
Y6V
FSig
CTd
27+t
BP
GO
9MFJ
QB
YWcj
AH
an+c
P8m
qm
aKd
xH/9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/65a
ef1
4973
0cb
2e1b
8f8e
c099
9d14
804/
5B06
D22
5/t5
1.28
85-1
5/e
35/c
130.
0.52
0.52
0/25
0222
80_2
0257
3693
7707
563_
4661
6256
4758
2412
800_
n.j
pg
FALS
EB
c-P
m5F
H8p
l15
1387
7075
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/e
43a3
536c
2829
4c5a
7cc6
f684
c97f
82d
/5B
2023
D6/
t51.
2885
-15/
e35
/250
2228
0_20
2573
6937
7075
63_4
6616
2564
7582
4128
00_n
.jp
gA
C-1
30J
Sup
er
He
rcu
les
NA
720
1080
2063
11.
97E+
0874
2063
1N
AN
AN
A
AC
oc0
Ft6a
1vW
jG6u
SB24
z9O
tBK
uD
jscV
l8yz
FaD
PSq
zwkV
ssQ
tU5m
z0p
KTB
oyX
TFR
bat
vUG
K2T
Mm
hIb
1ol4
PX
+Vb
drc
bb
cMTn
JOfX
k8f/
qrk
WP
NG
44rN
xua3
Oh
v70I
uEO
Wb
j/d
9c+/
YVl2
bM
0gX
soY4
/x9a
oZp
Q5X
kEg+
1CjZ
WQ
r3N
W6m
8scf
eP
T6e
v+FU
vtb
egq
sWLc
k5+t
PC
Aiq
SsI/
9k=
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
12b
cb5b
a076
014a
e33
e5a
bd
df6
a300
3/5B
157F
15/t
51.2
885-
15/e
35/c
130.
0.52
0.52
0/25
0153
95_1
3856
0170
1942
66_3
4123
4045
7131
2128
00_n
.jp
gFA
LSE
Bc7
vIC
Dn
3rj
1513
7929
36h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/dcc
052e
bf7
e55
22b
b83
f77e
09c4
e30
33/5
B1D
9657
/t51
.288
5-15
/e35
/250
1539
5_13
8560
1701
9426
6_34
1234
0457
1312
1280
0_n
.jp
gA
Re
pu
bli
c o
f Si
nga
po
re
NA
720
1080
1537
91.
97E+
0850
1537
9N
AN
AN
A
AC
oYt
Bak
CU
yS5j
ibae
SOvt
TDfJ
jjg9
vrTE
T7aj
wrc
Ag4
64rM
lnlY
bQ
eC
ck5H
4j6V
atH
8td
kYM
jHk+
g/lg
fWgY
5gC
xT07
/wB
Kj8
qrK
Wkg
bcS
F/U
/0H
86se
WK
Yh7K
r43A
Njp
kU4K
n90
fkK
KK
QD
EgjR
iyq
AT1
P+e
n4V
N06
UU
UD
GM
Sai5
oo
oEf
/Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/0ce
d0f
8f2f
74d
132f
3a91
9071
f80d
615/
5B11
7B43
/t51
.288
5-15
/e35
/c16
5.0.
449.
449/
2501
3816
_176
5587
3626
6158
_224
1718
3400
3056
2304
_n.j
pg
FALS
EB
c5K
SnrH
oX
715
1370
6515
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
17d
684e
a19b
3051
456e
7f23
b3c
5322
8/5B
20FB
AD
/t51
.288
5-15
/e35
/250
1381
6_17
6558
7362
6615
8_22
4171
8340
0305
6230
4_n
.jp
gA
C-1
30J
Sup
er
He
rcu
les
NA
621
1080
2005
71.
97E+
0837
2005
7N
AN
AN
A
AC
oax
aKK
K9E
5wo
oo
oA
Q0l
LSV
Ix4I
pC
RU
JpK
h1P
IrlJ
CTj
dg4
9ecU
gfP
akB
PTt
SkV
Ck9
7jsh
S1Jm
mgU
Yo5g
sf/9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/7c0
a15f
664d
710f
d67
046e
b21
0a4b
1f0/
5B15
03E8
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c19
7.0.
685.
685/
2502
5637
_192
7149
5461
8162
_679
2298
1063
9110
1440
_n.j
pg
FALS
EB
c2n
Hl3
H0c
S15
1362
0966
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
1f77
fad
f07e
9140
97b
01c0
a3ca
4aa5
3/5B
0BA
2D0/
t51.
2885
-15/
e35
/250
2563
7_19
2714
9546
1816
2_67
9229
8106
3911
0144
0_n
.jp
gA
B-1
B L
ance
r fr
om
N
A68
510
8018
784
1.97
E+08
5518
784
NA
NA
NA
AC
oX
sf2T
a9SD
/wB
9Gj+
y7P
OM
HP
8AvG
q6q
R1x
+dLu
VO
S3T2
rG/m
dH
I+xY
/sm
09D
/30a
T+yr
TsC
f8A
gRq
ATl
hx0
/z2p
8bgL
yT+B
xj0x
6U9S
Gkh
/wD
ZVp
/dP
/fR
qP
8As6
09D
/30a
iec9
Bn
H15
qru
Pq
fzp
2fcm
67FM
3hP
tSi4
B6n
NFF
VZA
5N7s
eLl
RyD
+hp
/2xc
Hn
r7G
iin
Ym5G
bh
T3/A
ENJ5
6+v6
Gii
gR//
2Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
2b27
bf8
536d
ca8e
203f
ad3e
7271
e91
d/5
A8C
88E2
/t51
.288
5-15
/e15
/c23
6.0.
607.
607/
2501
2033
_201
7962
0984
8685
1_86
0056
6405
5841
2595
2_n
.jp
gTR
UE
Bcu
pjo
2Hko
Y15
1335
4495
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/3
2a42
9843
66a9
9865
87d
925c
c26e
2742
/5A
8C80
FA/t
51.2
885-
15/s
1080
x108
0/e
15/f
r/25
0120
33_2
0179
6209
8486
851_
8600
5664
0558
4125
952_
n.j
pg
C-1
30J
airc
raft
fro
m t
he
54
602
607
1080
1477
11.
97E+
0874
1477
1N
AN
AN
A
AC
of0
VM
Lcb
xx1p
Xe
3Q4Z
ufz
/AJV
zyO
zZU
c7sf
p0p
Wb
pn
7wG
Djp
/+v1
96Yr
HTB
IgM
5GD
37U
h8o
Dd
kY9e
Mfn
WN
cOq
xqm
SzYB
Po
P6V
Vaf
Efl9
icn
3qSj
oSF
xntV
fzYv
7y/m
KaI
wU
Ayw
Ugc
e2O
lQfY
4fe
i4G
P93
Pak
DY/
Gm
PIG
5457
DP
HG
e4/
Ck3
fT9f
/iao
RK
WN
TEEK
AvO
V3H
HU
eu
e/G
KrL
83I7
/wB
Km
WV
UQ
g8M
SMN
6DB
yDw
eO
2O+f
akB
t25U
xqR
gZH
P17
1Jis
WD
UB
GgT
AO
O+S
Ovb
7p6U
p1c
DjZ
/wC
Pf/
Wp
DP
/Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/695
b91
022d
2c8a
a565
5a32
1132
fad
510/
5B0D
7634
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c13
5.0.
810.
810/
2484
5910
_170
0086
1267
2829
6_83
0840
4686
9525
9545
6_n
.jp
gFA
LSE
Bcs
DM
laH
3zx
1513
2665
88h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/df0
114d
b7f
2e78
b57
d0a
2a99
7d8a
a102
/5B
1267
B1/
t51.
2885
-15/
e35
/248
4591
0_17
0008
6126
7282
96_8
3084
0468
6952
5954
56_n
.jp
gG
en
era
l Jo
sep
h L
. Le
ngy
el,
@ch
iefn
gb, i
s o
n t
he
gro
un
d in
Cal
ifo
rnia
to
ass
ess
Nat
ion
al G
uar
d s
up
po
rt t
o c
ivil
au
tho
riti
es
resp
on
din
g to
wil
dfi
res.
#w
ild
fire
#ca
lifo
rnia
#ca
lfir
es
#nat
ion
algu
ard
#R
ep
ost
@ch
iefn
gb @
us.
nat
ion
algu
ard
NA
810
1080
1365
31.
97E+
0845
1365
3N
AN
AN
A
AC
ob
oR
hcA
Fcke
h/X
61ZA
Q9F
y3o
TzU
cZxx
x+P
+c1L
5pU
fKA
T+V
Aw
LSK
cbB
wP
Un
iph
uIy
Au
fqf8
/561
BK
yYB
ct8w
6Dq
Kjk
7Dn
GO
xxxQ
Bb
JGP
ugn
61m
C1b
1H5V
at0V
D35
9TV
jj3o
sBn
Kc/
WrA
UYy
xyB
6E1T
A+U
0KxH
50C
LDtv
YOm
Me
hz0
7Cm
Zcn
lhz7
Uo
5/H
rSK
fmx/
nrQ
MV
gAf8
P8A
PN
Lvb
3/O
mv1
xUR
UU
CP
/Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/8b
1dcb
b70
ca06
ac79
5d21
877f
97e
d89
d/5
B08
2264
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c18
1.0.
718.
718/
2500
9897
_194
0496
3795
3980
9_44
9012
3789
5221
7395
2_n
.jp
gFA
LSE
Bcp
xDu
QH
EiH
1513
1899
69h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/615
e8a
35cf
3f44
a6f0
8190
9f30
6361
4b/5
B1A
0A87
/t51
.288
5-15
/e35
/250
0989
7_19
4049
6379
5398
09_4
4901
2378
9522
1739
52_n
.jp
gA
irm
en
fro
m t
he
820
th
NA
718
1080
2043
21.
97E+
0874
2043
2N
AN
AN
A
AC
ob
1hlu
/wC
VSA
SDis
kaU
wP
7uQ
j65/
oar
TSSx
fujJ
5nP
OC
cew
56e
/XN
aNkW
Lmq
DzF
EYZd
ykFl
zyc8
AD
1q1P
dLC
oV
MM
VIB
9Bjq
M+v
06V
gPI8
43sc
tnk9
On
oB
2Ap
VH
AA
+6O
QR
/P/P
41N
yrF6
S+ft
wfU
f4d
Pzq
r9vl
/vD
8v/r
VE2
1h1x
UP
lH1F
LmH
Y03v
5J8o
g2r3
PfH
ue
nP
oK
yTK
Sem
ee
P89
/xq
/8A
chA
XjP
Wq
UQ
G3P
fP86
bEi
YZ+7
+WP
ep
hFg
YJw
P89
am2h
QM
VEx
OcV
G5e
xCy5
GR
UO
33q
Rzj
gdq
q7j
QB
//9k
=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/a46
40e
92a9
cbf3
200b
108a
d0a
e1e
97c2
/5B
1187
15/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
181.
0.71
7.71
7/25
0059
40_3
7178
9406
6132
94_8
6343
6946
4650
9568
0_n
.jp
gFA
LSE
Bcn
Ih-v
gtu
X15
1310
1612
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
9952
3270
910d
f3e
292d
8ea6
5bb
8f82
1/5B
068D
A8/
t51.
2885
-15/
e35
/250
0594
0_37
1789
4066
1329
4_86
3436
9464
6509
5680
_n.j
pg
A s
pe
cial
mis
sio
ns
NA
717
1080
1696
61.
97E+
0825
1696
6N
AN
AN
A
AC
ocw
VV
Sdp
4x3/
xqY2
jAge
tRm
Mty
Oo
6+/w
Dn
vWn
ZtgY
J3Z7
elJ
3Q1Z
me
YCD
io2i
xXQ
KiB
xuw
Rn
JB64
qlf
Mh
b5e
ntU
qX
Qb
Rjl
aTb
VlY
93X
j60e
X9K
sk0m
Xcf
lTH
I5IP
Tv9M
np
Ub
xNC
2+P
5QfX
oM
+5q
D7Z
KB
nca
c8ry
qN
xJyR
U3b
3Ksl
sOcM
x3Ll
s9/w
Co
9vw
FTe
UG
4KEE
An
vyR
2696
akjR
kAd
v6d
Ku
Pcu
V7A
jnP
es3
J7F2
KD
gLk+
X16
Ag8
cck8
9c9K
AYu
8b/n
VlL
tz1w
fwp
fPb
0H5U
ud
rovv
Dlv
1P/9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/23b
4a04
a299
f5a9
8ac1
ff17
49a7
d7f
d5/
5B10
FB31
/t51
.288
5-15
/e35
/c13
0.0.
520.
520/
2501
8698
_309
7154
1953
0272
_830
7231
6841
3941
76_n
.jp
gFA
LSE
Bck
nO
STH
ue
o15
1301
7041
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/f
5861
f67c
46c6
9d7f
be
f4b
bb
30a8
4c2c
/5B
1E22
34/t
51.2
885-
15/e
35/2
5018
698_
3097
1541
9530
272_
8307
2316
8413
9417
6_n
.jp
gA
Ne
w J
ers
ey
Air
N
A72
010
8021
497
1.97
E+08
5121
497
NA
NA
NA
AC
ob
ga1S
fom
wli
hb
+EH
oe
h5I
Ptg
+tB
tLfO
Aj5
OcY
B5A
6kd
citV
rWO
SLyT
naM
dO
uQ
c9fz
qw
qK
CC
Bgq
No
9hxx
+g96
ZBzM
1tG
uA
gIJI
yG4I
zxg+
lJN
aIM
9Vw
cE7S
Rn
0yB
ip5V
aWU
kDh
nO
D9M
9fTF
arxC
SLaO
M85
985z
+lA
XM
e2t
kT5m
Hm
Kp
wyl
SGG
QSM
Dq
c+p
xilM
1l/w
A8z
/n8a
1PJA
l87J
zjp
2Po
fwFU
3sYX
Yt0y
Sfzo
C5s
qB
S5xU
KE1
MR
QB
hO
Xh
m24
JGW
Ixn
BLY
6+vf
861Y
1KIF
PU
Dn
60yZ
QX
jPd
S2P
++aT
caA
BiT
Vb
Bq
dia
r7jT
A//
Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/069
7473
141d
3935
3629
6fd
8a1a
2ed
8d0/
5B13
D1F
1/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
181.
0.71
8.71
8/25
0068
41_1
1524
8768
1551
656_
6445
4664
2765
3619
712_
n.j
pg
FALS
EB
cc1z
OYn
Arp
1512
7562
48h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/a9e
1a1c
26af
8c2d
1217
262b
6f5c
64b
4f/5
B03
3112
/t51
.288
5-15
/e35
/250
0684
1_11
5248
7681
5516
56_6
4454
6642
7653
6197
12_n
.jp
gA
U.S
. Air
Fo
rce
A-1
0 N
A71
810
8023
479
1.97
E+08
100
2347
9N
AN
AN
A
AC
ob
0VFS
Vkj
UsA
/u2G
PX
8vSm
/wB
sKO
qH
8/8A
61A
jVd
1UgE
gFjw
PW
hu
BW
LLq
O/b
IAD
hiA
O/4
/zG
O9T
DV
BIO
EY+u
OaQ
yG4v
/LYq
O1V
P7Q
NV
Lhw
7lh
xn1q
DN
MD
olv
Pr+
Zp/2
z2z9
axQ
xp29
vWgR
pG
VN
ykK
BtJ
OA
B1P
ekn
nij
ban
3QB
gAcZ
PJ5
HfO
f5V
Q3n
I5p
p9K
YFp
Jo+Q
CT6
YAyP
zpP
NT1
b8l
qjR
QB
//Z
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
636f
df2
3b9b
2ae
30fb
7d2a
2977
c33c
d/5
B4D
98FB
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c18
3.0.
714.
714/
2501
0896
_147
7992
7253
6376
_197
0345
9209
6904
8064
_n.j
pg
FALS
EB
caTA
O8n
lJw
1512
6708
95h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/1a5
e7a
d53
a9d
6d43
de
7bf6
ba4
e3d
862e
/5B
108B
6C/t
51.2
885-
15/e
35/2
5010
896_
1477
9927
2536
376_
1970
3459
2096
9048
064_
n.j
pg
A U
-2 D
rago
n L
ady
lan
ds
NA
714
1080
2688
71.
97E+
0812
526
887
NA
NA
NA
AC
oc5
tVLH
A5J
4Fay
wiB
Mh
Szd
6j0e
NZL
gbu
wJ/
Hp
/Wtu
4jC
sdve
ue
pLW
3Qw
qX
+Rjr
dH
dsY
HH
TH+f
TrxU
YmV
Mo
RlS
frir
Fw2F
Ct9
4DA
PX
I7Ee
316V
ntC
wG
7BxQ
kn5E
JJ76
Ec8Y
Q5X
oah
qd
1yD
7dag
rZb
HR
HYn
tpjD
IHH
b+R
rZaf
ec1
gYq
xC56
dq
icb
6mc4
31Lo
CM
Mty
e34
dam
eYB
QO
gB59
PU
1RZi
MA
Usx
wn
Hq
az5b
2Mrb
IrTv
uYn
1qG
loxX
QtD
qSs
rH//
Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/20d
d02
c774
761e
0ab
4b9b
9787
ca29
a41/
5B01
8507
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c18
0.0.
720.
720/
2484
5460
_525
4056
4448
2277
_250
8295
7115
7393
4080
_n.j
pg
FALS
EB
cXtY
z9H
LTm
1512
5840
65h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/842
d64
c717
d34
efe
bb
94b
415d
fa2a
5a5/
5B20
0D07
/t51
.288
5-15
/e35
/248
4546
0_52
5405
6444
8227
7_25
0829
5711
5739
3408
0_n
.jp
gM
em
be
rs o
f th
e
NA
720
1080
1693
21.
97E+
0863
1693
2N
AN
AN
A
AC
ocQ
oX
4Kg/
U/w
CA
NV
DII
iVD
4x2H
zAfi
asX
pYK
Ao
OD
1I7V
nrG
QSW
Ugf
Q/w
Ce
a0ky
Ei0l
6h4b
P1x
/ID
+tW
Q8Z
6ED
61n
+VK
wA
Cn
j2p
4sp
m68
fU0l
Jg0i
8So
7jH
qC
KYW
bP
Ck+
+RV
QW
DFt
rsB
xkd
8/Tp
0qx/
Z6f3
mq
rye
wrJ
Giy
Bxt
bkH
tUd
y4V
QvX
ed
vP8A
n8v
erx
yf/w
BQ
/wA
KZJ
CJA
N3b
6eo
Pp
7Vn
zFcp
Cp
Q/K
pH
HG
M8j
H+F
LikW
0SJt
6k55
64xz
17V
KB
n/8
AU
P8A
Cjm
sHK
Qsg
Ygn
kryP
Y0Yq
x5Q
P+R
/hR
5Q/z
j/C
jnH
yn//
Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/58f
2dcd
c32e
8255
0605
d5b
59ac
5035
48/5
B05
B6E
D/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
180.
0.72
0.72
0/24
8457
40_9
6264
6540
5665
40_1
8613
4906
1269
9136
00_n
.jp
gFA
LSE
BcV
HEw
1ng8
u15
1249
6869
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/6
462b
5fae
6c58
fc1c
67fb
de
b9c
57f1
e8/
5B1B
DA
ED/t
51.2
885-
15/e
35/2
4845
740_
9626
4654
0566
540_
1861
3490
6126
9913
600_
n.j
pg
Air
me
n f
rom
th
e 2
6th
N
A72
010
8016
524
1.97
E+08
2916
524
NA
NA
NA
NA
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/4
eb
f92c
cfa9
ec4
64c0
bf3
12ce
c765
d5e
/5B
16D
FFE/
t51.
2885
-15/
e35
/c12
0.0.
366.
366/
2432
7623
_159
3895
8814
6133
_436
7234
7163
7020
6720
_n.j
pg
FALS
EB
cSle
Fmn
nyZ
1512
4121
42h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/be
9034
784f
256c
b69
8034
b41
3976
9245
/5B
2038
B7/
t51.
2885
-15/
e35
/243
2762
3_15
9389
5881
4613
3_43
6723
4716
3702
0672
0_n
.jp
g43
2nd
Air
craf
t N
A38
664
013
993
1.97
E+08
2013
993
NA
NA
NA
AC
od
6QsA
Mn
oK
iiu
Ypv8
AV
srH
0B5/
LrSu
Mq
QO
CQ
f5V
hW
9uV
lDIQ
Cgy
xYZx
9Dkd
f8A
69H
kNK
50d
ULx
n+4
p25
Gfr
/Wp
/tC
AgE
4JO
Bn
v9D
/Ko
L8xq
m6T
PH
Tbw
T3x9
PW
ize
iFd
LfYg
0yR
yXR
s4TH
X1P
pm
tau
Uad
lHJI
JGM
D88
fnTh
cnH
/AN
eu
hU
W1v
qZO
ok9
tB0u
ps/
3h8p
x8g/
yM/w
Aq
LW9a
AkE
F1Yg
YH44
Iz2I
Izn
vWN
Gu
4k5w
Rx+
dW
kch
WO
fugf
zIrl
Ni9
c36S
n5c
qy5
Aw
Mke
/Xj8
Kry
zySD
zHB
8sY2
57n
+ucc
mp
bf9
2ygf
xck9
yev/
wB
bjF
Ou
QD
CR
joQ
B7b
up
+vFb
wkl
JWW
r0Ik
rrV
mV
JNls
+hq
XzP
b9K
YUH
SmG
XH
GP
5f4V
27b
mFr
7H//
2Q==
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/b
4078
dc6
b0c
bb
6174
e46
bd
3225
360f
f5/5
B20
4CA
8/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
157.
0.76
5.76
5/24
3272
27_1
9229
3413
1304
820_
4486
8016
0225
3946
880_
n.j
pg
FALS
EB
cK8h
Iqn
m3J
1512
1557
90h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/cb
7865
3b1b
11d
6d48
a11e
3b3e
91a8
7d1/
5B09
610E
/t51
.288
5-15
/e35
/243
2722
7_19
2293
4131
3048
20_4
4868
0160
2253
9468
80_n
.jp
gSe
cre
tary
of
the
Air
Fo
rce
NA
765
1080
2491
71.
97E+
0819
824
917
NA
NA
NA
AC
oa0
47d
U9f
zp7Q
KR
kfK
fUE1
S1C
+Nq
g2Y3
Pn
Ge
wH
fHfr
9Kp
W6e
fIsi
Oc8
PIu
c49t
3Gcn
+E9O
fSm
I2lK
INp
bj2
5qTE
TfN
wcV
F58a
Hac
Z+gq
wN
pG
Vxg
+lFw
Gtt
AyO
Pp
UX
/AH
1Sn
cR/s
io/t
JHr+
lIZz
F7N
50q
g9FA
BO
fxP
6Vrf
a0ij
Cw
qB
n07
e59
T9az
Ao
Ms+
QD
hW
x7d
Kji
Py/
jTEW
Q24
5PX
rnN
aVvc
lRgn
j9aw
ZOh
qzb
EkD
PP
FDG
b32
vbw
PzP
Sqxv
T6D
9ap
MTv
b6j
+Qp
1Aj/
/2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/111
567f
2374
94d
0e36
1169
7660
672d
4d/5
B21
9500
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c20
3.0.
673.
673/
2412
5341
_174
3755
2659
3437
4_66
5037
8943
6836
2496
_n.j
pg
FALS
EB
cIfn
c4H
Xjx
1512
0735
28h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/074
bb
ff6a
eac
bd
3c27
3d4e
860e
5e2e
1f/5
B15
03B
7/t5
1.28
85-1
5/e
35/2
4125
341_
1743
7552
6593
4374
_665
0378
9436
8362
496_
n.j
pg
U.S
. Air
Fo
rce
Sta
ff S
gt.
NA
673
1080
1654
91.
97E+
0852
1654
9N
AN
AN
A
AC
ofl
sofL
jzjB
71H
FZu
UZ2
Klz
kkD
ofQ
+xzj
kfrV
+35Q
Y5G
MfX
HW
pSF
X6m
uN
vVn
Xto
cc7E
MfU
E81d
snd
mU
YBC
kDJJ
zg8Y
A6c
+ntm
oJ1
VJC
MH
IPp
xj8O
p9f
8a2L
e2G
Ud
Bh
AM
8jB
yf5/
wD
6sH
FbTa
5fU
yitX
rsTz
xBga
xzG
o9K
6OV
Mis
do
JMn
5l6+
lYw
fLd
Nm
z95L
S5sw
4ID
D0F
Nu
LpYU
3MM
qTj
8aE+
Ub
e2K
guo
Gn
i2q
QC
p3D
PTj
1/C
has
iS6n
PX
Mxm
be
Rgd
RW
3p9+
bgF
JeW
XG
CO
49/c
f1rn
TIW
4Hb
+laG
m7h
KZX
+7w
hI9
W+7
x6ZH
b+V
dM
o3V
jBP
XU
6N24
rMaK
Ukk
NgZ
q67
dq
r7q
wSa
OlW
sf//
Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/c42
8aa8
bd
fda5
42c6
3f74
4643
0a17
e37
/5B
196B
6E/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
135.
0.81
0.81
0/24
1262
77_6
8233
8455
2230
83_4
4212
1614
1679
0016
_n.j
pg
FALS
EB
cFtU
thn
FJW
1511
9800
52h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/1e
e69
1b27
faa4
0d5e
7c0d
6dd
0ad
6986
c/5B
1239
B8/
t51.
2885
-15/
e35
/241
2627
7_68
2338
4552
2308
3_44
2121
6141
6790
016_
n.j
pg
Fire
figh
ters
wit
h t
he
1st
N
A81
010
8019
033
1.97
E+08
5219
033
NA
NA
NA
AC
ocx
YUzX
QW
qb
RX
Pxy
ba2
re6X
bik
wN
yPB
qyK
y45l
B6g
Z9+t
Xkm
Dd
KS0
GSs
Mjm
s8xL
mp
L65M
EW5f
vFlU
Z9z/
hm
kZh
k0m
Bw
4ap
0k21
Vp
w61
YjV
hu
AD
uJH
Pr6
f561
etp
lhG
FO4d
hu
U//
XxW
ATT
T1p
Ab
2o3P
mP
FHn
+IM
fzA
H9a
le9j
DEZ
HU
9655
hsw
QSK
ho
sB//
Zh
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/9ae
bcf
b6f
cd82
5253
bd
1ec9
4fe
0048
e5/
5B0A
EC32
/t51
.288
5-15
/s64
0x64
0/sh
0.08
/e35
/c18
0.0.
720.
720/
2412
6608
_146
8316
2596
2737
_371
7121
8848
9505
1776
_n.j
pg
FALS
EB
cDQ
E8YH
VX
X15
1189
7609
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/8
9217
1659
1d09
ca77
23d
ac5f
70e
e36
e8/
5B12
3732
/t51
.288
5-15
/e35
/241
2660
8_14
6831
6259
6273
7_37
1712
1884
8950
5177
6_n
.jp
gA
n A
irm
an f
rom
th
e 9
2nd
N
A72
010
8014
289
1.97
E+08
3014
289
NA
NA
NA
AC
ok5
ulx
QD
jpU
qu
Oh
pgC
jHT0
5pN
mB
9alD
D6Y
oaU
tycf
Qcf
pSA
rlSK
bT2
ctTK
YBR
RR
QA
8N2x
T9h
7d+a
hq
Zj8o
oA
dH
Go
z5m
Rkc
Htn
tn29
cVC
aAxF
JuN
AC
UU
UU
AFL
njF
FFA
CU
UU
UA
f/2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/bb
410e
9d99
9f94
b84
dd
ad12
92c2
b54
78/5
B02
9244
/t51
.288
5-15
/e35
/c39
.0.5
22.5
22/2
4125
613_
1426
8682
3120
030_
6672
5389
9924
6028
80_n
.jp
gFA
LSE
BcA
nn
e2n
are
1511
8092
88h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/2d
57fd
6aaf
9676
a0b
7e20
25ab
5c2c
970/
5B24
22ED
/t51
.288
5-15
/e35
/241
2561
3_14
2686
8231
2003
0_66
7253
8999
2460
2880
_n.j
pg
A K
C-1
0 Ex
ten
de
r fr
om
N
A55
664
023
612
1.97
E+08
4923
612
NA
NA
NA
AC
oaY
oq
ZcZq
Oo
8MM
4xz6
0hl/
eA
Md
6hLG
qLP
J+X
pU
iSlh
yCD
QFi
YuaN
wq
ItTc
0wIB
Oac
bjt
VEU
lSM
ue
djk
U03
War
1Ee
tAFs
zCp
N4q
gKlF
AH
//2Q
==h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/2b
fe4e
d11
3866
af99
d77
8be
f729
0cb
bf/
5B01
FAB
4/t5
1.28
85-1
5/s6
40x6
40/s
h0.
08/e
35/c
198.
0.68
3.68
3/23
7341
42_1
5351
1155
3271
630_
1944
9237
3466
6117
120_
n.j
pg
FALS
EB
bzX
dao
HK
iN15
1136
4609
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/9
125c
c270
c8c7
b41
3145
ec1
9e28
51cd
7/5B
051D
D7/
t51.
2885
-15/
e35
/237
3414
2_15
3511
1553
2716
30_1
9449
2373
4666
1171
20_n
.jp
gA
U.S
. Air
Fo
rce
HH
-60
NA
683
1080
2246
71.
97E+
0862
2246
7N
AN
AN
A
AC
ocu
5Sp
F2n
pV
MC
nZx
1NcD
Z2tF
3cFx
wTn
0Ap
xizz
n86
qK
vv+t
WER
e7C
snK
+lvx
sZN
W6/
gBip
m2r
YiQ
jOcj
61Fi
H1/
UU
OD
XV
L5iU
vX7j
kftE
h/i
pP
NkP
VjU
K06
vVsu
xld
9yw
J3/v
Gn
rMw
75/C
qtG
ahxR
op
M1F
mO
08n
9KrG
Q+t
QK
xwaZ
msl
C1z
Ry2
P/9
k=h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/dd
4dc9
f164
141f
b71
a692
3ae
8bf7
21e
a/5B
4E61
54/t
51.2
885-
15/s
640x
640/
sh0.
08/e
35/c
180.
0.72
0.72
0/23
8245
02_5
3687
9203
3276
83_3
0106
7452
7198
7732
48_n
.jp
gFA
LSE
Bb
xDrK
YHd
NN
1511
2871
27h
ttp
s://
sco
nte
nt-
ort
2-2.
cdn
inst
agra
m.c
om
/vp
/67d
7a0d
8572
752f
ed
a5f0
1b69
470f
3b0/
5B21
EE54
/t51
.288
5-15
/e35
/238
2450
2_53
6879
2033
2768
3_30
1067
4527
1987
7324
8_n
.jp
gA
U.S
. Air
Fo
rce
C-5
M
NA
720
1080
2517
31.
97E+
0885
2517
3N
AN
AN
A
AC
oc5
9RS4
yaFp
rHm
gCcj
BxU
X8V
IG9a
UH
JoA
GFQ
1O1Q
0ASK
aQgZ
54p
BQ
1AC
Cn
r1p
gpw
60A
PY1
DTy
ajo
A//
9k=
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/7
7b2e
3bce
ac8a
8083
6ae
5008
2d32
222a
/5B
05F4
6C/t
51.2
885-
15/e
35/c
129.
0.52
1.52
1/23
7346
52_1
3342
8990
6518
78_1
5826
0706
2444
0156
16_n
.jp
gFA
LSE
Bb
uff
ASn
uo
e15
1120
1044
htt
ps:
//sc
on
ten
t-o
rt2-
2.cd
nin
stag
ram
.co
m/v
p/2
6bb
2369
30d
2ace
f5aa
57c8
89f5
60d
35/5
B08
071F
/t51
.288
5-15
/e35
/237
3465
2_13
3428
9906
5187
8_15
8260
7062
4440
1561
6_n
.jp
gU
.S. A
ir F
orc
e C
apt.
Jam
es
NA
721
1080
2281
91.
97E+
0849
2281
9N
AN
AN
A
61
Appendix F
id text created_at owner.id owner.profile_pic_urlowner.username
17902716826081000 Nice!! 1516644761 1257719479 https://scontent-ort2-2.cdninstagram.com/vp/6e9dc1d18060a803c403d0fa1284e47e/5B1531BB/t51.2885-19/s150x150/24126366_139165150008018_2795172155120156672_n.jpgsoph_ontwowheels
17892879493145000 <f0><U+009F><U+0091><U+008C><f0><U+009F><U+008F><U+00BB>1516644775 306180651 https://scontent-ort2-2.cdninstagram.com/vp/2a1d9e70f9c5472848c18ff719f76afd/5B49ABE9/t51.2885-19/s150x150/27890965_2099072973656754_7303907244821708800_n.jpgsallykoonin_123
17920219525007200 Good girl! <U+2764><f0><U+009F><U+0092><U+00AA>1516644783 31383965 https://instagram.fagr1-1.fna.fbcdn.net/vp/3230896e49952035c4a21d078561d30f/5B1DB27A/t51.2885-19/11906329_960233084022564_1448528159_a.jpgcc_the_dj
17846245039238900 ah 1516644792 1581502399 https://scontent-ort2-2.cdninstagram.com/vp/474effd3afd430c302057e586bd8a3f7/5B4AE6C8/t51.2885-19/s150x150/11856685_981735955211282_219583063_a.jpgbig_candycane_lane_madeosteel
17929182385004300 Get it. 1516644802 3016165931 https://scontent-ort2-2.cdninstagram.com/vp/3d7412d5dc34b278dc9edf938dd40c9d/5B1AB5F5/t51.2885-19/s150x150/23498420_146923939275945_5315202627879829504_n.jpgdennisgivens
17859767317219600 <f0><U+009F><U+0091><U+008F><f0><U+009F><U+0091><U+008F><f0><U+009F><U+0091><U+008F><f0><U+009F><U+0091><U+008D>1516644827 3922217374 https://scontent-ort2-2.cdninstagram.com/vp/27fa9aed991bb98fe03072086ae170ff/5B1139A4/t51.2885-19/s150x150/24846014_362064277586536_2437640544617234432_n.jpgsam.deylami4054
17919995560054000 its a timex expedition moment1516644852 1581502399 https://scontent-ort2-2.cdninstagram.com/vp/474effd3afd430c302057e586bd8a3f7/5B4AE6C8/t51.2885-19/s150x150/11856685_981735955211282_219583063_a.jpgbig_candycane_lane_madeosteel
17902404370091000 @anna.ekstromh kan inte du va såhär cool?1516644867 2222529160 https://scontent-ort2-2.cdninstagram.com/vp/9b6c447f0cb7df0c14a7f5df3fc71a8f/5B1D70E5/t51.2885-19/s150x150/27877706_204447570291007_9108200985573982208_n.jpgville360
17846218849241100 Nice 1516644876 5713619573 https://scontent-ort2-2.cdninstagram.com/vp/f40c811ddb5b8e07f28f62efd407e93e/5B28DC9C/t51.2885-19/s150x150/26319515_528530067515824_1373147541847146496_n.jpgairforce.ir
17846242351233800 That is awesome... My daughter wants to be a pilot in the Air Force.. You are such and inspiration. @usairforce1516644878 270062153 https://scontent-ort2-2.cdninstagram.com/vp/7de4cfd05e7b680cbc833706ca1b6d7f/5B1B554C/t51.2885-19/s150x150/18947843_650353381831190_2087312547752443904_a.jpgrozyp77
17892406873161800 Eyebrows on fleek 1516645059 33344140 https://scontent-ort2-2.cdninstagram.com/vp/e2796b68f027501ec5d114aec8dfb00d/5B2BD081/t51.2885-19/s150x150/27879767_295904594267297_3289743914201776128_n.jpgtreybrah_
17907310702129400 <f0><U+009F><U+0091><U+008A>1516645111 39121918 https://scontent-ort2-2.cdninstagram.com/vp/c6a059f1da8a1f0b88436c60514ae6fd/5B1E50D8/t51.2885-19/s150x150/26392794_2001319253490120_9011118206240161792_n.jpgmekala333
17896066945138500 Shaw AFB represent! <f0><U+009F><U+0087><U+00BA><f0><U+009F><U+0087><U+00B8>1516645119 34501017 https://scontent-ort2-2.cdninstagram.com/vp/3f38c8630a1c294bb62a54d39a48e810/5B4C1BDB/t51.2885-19/s150x150/26866190_499624393770957_8844942846361337856_n.jpgchrisastro
17923311298058600 SALUTE!! 1516645150 6481124 https://scontent-ort2-2.cdninstagram.com/vp/b86c08814740bdb1545218bba73af9a5/5B4A2C21/t51.2885-19/11881824_986618044721646_1650377176_a.jpgpaulagoble
17919239926045500 @idcvickyy u can be her.1516645242 415630583 https://scontent-ort2-2.cdninstagram.com/vp/118d8643ead538c26431286035861f40/5B4BFAF0/t51.2885-19/s150x150/27893291_168703120427388_2988161910655942656_n.jpgjoha_balay
17893049878150400 YEA! Viper pilots rule1516645267 3266740512 https://scontent-ort2-2.cdninstagram.com/vp/ffb4cf6739f26fd4447ffd38ae372ec7/5B11AAA9/t51.2885-19/s150x150/26272222_570047826669746_6501235328156696576_n.jpgonyxreaper
17902779208084300 <f0><U+009F><U+00A4><U+0098><f0><U+009F><U+008F><U+00BB>1516645297 210947008 https://scontent-ort2-2.cdninstagram.com/vp/ea86b958e8150aa3ae4e12fbb4aba1cf/5B2A2F79/t51.2885-19/s150x150/16585096_476771682711740_4484955204403331072_a.jpgcrazy_gpigz
17846326054239800 Ville, jag är 14<f0><U+009F><U+0098><U+0082> @ville3601516645301 857171570 https://scontent-ort2-2.cdninstagram.com/vp/72b0c949a73b5f8a74158c20e5dd8fb9/5B1766AC/t51.2885-19/s150x150/25005578_663682734019309_7577762945254293504_n.jpganna.ekstromh
17907677278112200 #goals<U+2764><U+FE0F><f0><U+009F><U+0092><U+00AF>1516645343 22737475 https://scontent-ort2-2.cdninstagram.com/vp/29278116f4d7753f08165c42beabb172/5B1613D8/t51.2885-19/s150x150/26181292_443533316063348_3256149732279451648_n.jpggiaxlana
17895607483134500 <f0><U+009F><U+0091><U+008F><f0><U+009F><U+008F><U+00BC><f0><U+009F><U+0091><U+008F><f0><U+009F><U+008F><U+00BC><f0><U+009F><U+0091><U+008F><f0><U+009F><U+008F><U+00BC><f0><U+009F><U+009A><U+0081><f0><U+009F><U+00A6><U+0085><U+26A1><U+FE0F><f0><U+009F><U+0092><U+0080><f0><U+009F><U+0087><U+00A7><f0><U+009F><U+0087><U+00B7>1516645352 4608014410 https://scontent-ort2-2.cdninstagram.com/vp/a1c7b6e6f49dd5314ac8d423363fe4a4/5B1DC428/t51.2885-19/s150x150/25026154_1862082484104283_7468195177027338240_n.jpgciopaer_to
17919464815010700 @amanduh0429 <f0><U+009F><U+0098><U+008E>1516645382 400073675 https://scontent-ort2-2.cdninstagram.com/vp/81b090a1da1636247301ecc73ee88c1d/5B0EEFD0/t51.2885-19/s150x150/22639157_180220312553059_2932803269723947008_n.jpgcacraig11
17879493088196800 <f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D>1516645436 4979572474 https://scontent-ort2-2.cdninstagram.com/vp/4a8b9b71c86d72d3e4b0013e5ce06e8c/5B120E93/t51.2885-19/s150x150/27891601_196679197587689_7702282456711299072_n.jpgjaymon3002
17919810205032800 <f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D>1516645436 4979572474 https://scontent-ort2-2.cdninstagram.com/vp/4a8b9b71c86d72d3e4b0013e5ce06e8c/5B120E93/t51.2885-19/s150x150/27891601_196679197587689_7702282456711299072_n.jpgjaymon3002
17894531890188700 Sickkkk 1516645519 1981970767 https://scontent-ort2-2.cdninstagram.com/vp/c0f6d68e43f75b28eeb8bae8465b3bcd/5B1A5341/t51.2885-19/s150x150/26227541_156133235038686_1302106944525303808_n.jpgjust.__tin___can_
17920278703051200 <U+2764><U+FE0F> my heart at Eglin <f0><U+009F><U+0087><U+00BA><f0><U+009F><U+0087><U+00B8> #USAFmom1516645582 285396048 https://scontent-ort2-2.cdninstagram.com/vp/7c705b3964588d105058a11211e14575/5B147527/t51.2885-19/s150x150/13397676_1792024331013142_2095304786_a.jpgj_yvette_l
17894926300130200 Can you get her for career day <f0><U+009F><U+0098><U+0082> @mo_elizabeth3 ?1516645687 295804343 https://scontent-ort2-2.cdninstagram.com/vp/f4e2c8eac6d31bcf283b75023d4b325c/5B10BC8C/t51.2885-19/s150x150/25008072_171693896897547_257306796304105472_n.jpglo_lights_
17902481515084500 NND PBSBHBJBLBZBXBCB Q RDGJKLCB RDGJKLCB RDGJKLCB RDGJKLCB RDGJKLCB RDGJKLCB RDGJKLCB RDGJKLCB RDGJKLCB RDGJKLCB RDGJKLCB1516645713 6330870614 https://scontent-ort2-2.cdninstagram.com/vp/1c5687ff2d0a685bd95181ef071ae865/5B048A09/t51.2885-19/s150x150/23164537_135653563756420_5548774266675134464_n.jpgnoa727
17906989870099700 Veeeeery niiiiiiiiiice<U+270C><U+270C><f0><U+009F><U+0092><U+0095><f0><U+009F><U+0092><U+0095><f0><U+009F><U+0098><U+0099><f0><U+009F><U+0098><U+0099><f0><U+009F><U+0098><U+009A><f0><U+009F><U+0098><U+009A><U+270C><U+270C><f0><U+009F><U+0092><U+009E><f0><U+009F><U+0092><U+009E><U+270C><U+270C><f0><U+009F><U+0092><U+0095><f0><U+009F><U+0092><U+0095><U+270C><U+270C>1516645731 3579379587 https://scontent-ort2-2.cdninstagram.com/vp/e00f983169e55b62e85a03b7edebe8e6/5B05688A/t51.2885-19/s150x150/27575031_171202726997569_2043483463058718720_n.jpgkoroush_imani
17879685745196500 @lo_lights_ I asked for a myriad of officers ranging from 2d lt to Capt1516645755 424550255 https://scontent-ort2-2.cdninstagram.com/vp/887e8a0f66d9425b7e723227b0ef19f2/5B04FF84/t51.2885-19/s150x150/18094555_254650541667515_2726418108192391168_a.jpgmo_elizabeth3
17920758868017800 @frankpaint 1516645776 46654786 https://scontent-ort2-2.cdninstagram.com/vp/9352930c4d18d4c8df7eb21f23538e7c/5B15859A/t51.2885-19/927393_855653307791887_527646871_a.jpglaurenpainter_
17915043367070100 @mo_elizabeth3 I just like that she's a female fighter from Shaw.1516645806 295804343 https://scontent-ort2-2.cdninstagram.com/vp/f4e2c8eac6d31bcf283b75023d4b325c/5B10BC8C/t51.2885-19/s150x150/25008072_171693896897547_257306796304105472_n.jpglo_lights_
17924786815057300 Found <f0><U+009F><U+0092><U+00A7><U+270A>1516645911 2266396868 https://scontent-ort2-2.cdninstagram.com/vp/69f1cae72e6271b6624cde188193c67a/5B1233E4/t51.2885-19/s150x150/27892042_2048175975420147_6924264451189243904_n.jpga.badhisutawan
17921940580013500 <f0><U+009F><U+0099><U+008C><f0><U+009F><U+008F><U+00BB> cool <f0><U+009F><U+0098><U+008E>1516645921 3514876709 https://scontent-ort2-2.cdninstagram.com/vp/e3fbfeb1641cfcbff13d7505b8b4aeb4/5B07B835/t51.2885-19/s150x150/21984976_1739498929416714_3158543980547801088_n.jpgfawzinelfarr
17907142621097000 Bravo Bravo Bravo <f0><U+009F><U+0098><U+0080> <f0><U+009F><U+0098><U+0080><f0><U+009F><U+0091><U+008F><f0><U+009F><U+0091><U+008F><U+270C><U+FE0F><U+270C><U+FE0F><f0><U+009F><U+0098><U+0098><f0><U+009F><U+0098><U+0098><f0><U+009F><U+008C><U+00B9><f0><U+009F><U+008C><U+00B9>1516645957 4755191572 https://scontent-ort2-2.cdninstagram.com/vp/de2d83fba16f557704ad88dccdd9e6c6/5B2AEE98/t51.2885-19/s150x150/26869666_199713793941897_8150894711356260352_n.jpgmiro.jaume
17893056922148200 Great picture , thank for your services... God bless you ..1516646062 4699730528 https://scontent-ort2-2.cdninstagram.com/vp/2de3ea6bd4ecef8b4e32c6038d7449e1/5B169D39/t51.2885-19/s150x150/28155962_2046093055672503_3645957326227111936_n.jpglmma2003
17904741163126700 You shot the @usairforce1516646228 3228699880 https://scontent-ort2-2.cdninstagram.com/vp/2643911ee8e1ddebfb46f2ee22ce264a/5B4A52F6/t51.2885-19/s150x150/27581997_1991437354512823_1111268654072201216_n.jpgian_t_wilson_
17905736824126400 Oh wait wrong picture1516646264 3228699880 https://scontent-ort2-2.cdninstagram.com/vp/2643911ee8e1ddebfb46f2ee22ce264a/5B4A52F6/t51.2885-19/s150x150/27581997_1991437354512823_1111268654072201216_n.jpgian_t_wilson_
17905311511124800 #CooL 1516646416 4285556838 https://scontent-ort2-2.cdninstagram.com/vp/6475f971cc2cdf3667ab612a380118fc/5B1E5BDF/t51.2885-19/s150x150/25021995_712730572270192_1895124379142979584_n.jpgmarlilny
17860723798209200 @jadentkorth Eglin hua1516646421 2064800284 https://scontent-ort2-2.cdninstagram.com/vp/8487ec6298d9a4bdf4433b63767b9438/5B0D9F47/t51.2885-19/s150x150/25018711_239437903262579_3841898226806423552_n.jpganthonyliz_
17920338988032700 @anthonyliz_ HUA 1516646464 246687008 https://scontent-ort2-2.cdninstagram.com/vp/a1217065a3f24d9db484713a73e19b81/5B1D98F5/t51.2885-19/s150x150/26152031_193338297912578_7166435164266954752_n.jpgjadentkorth
17893257619192700 BEAUTY AND STRENGTH! <f0><U+009F><U+0098><U+0087><f0><U+009F><U+008C><U+00B9><f0><U+009F><U+0092><U+008B><f0><U+009F><U+008E><U+00B5>1516646513 4871254374 https://scontent-ort2-2.cdninstagram.com/vp/8d2167e74606ec1265d2e4df8e0a361b/5B297DB2/t51.2885-19/s150x150/17439306_283536282059742_1794652896382418944_a.jpgjjascosmos
17920623730011200 @jamaican_johnny @walker_polsinelli @kylematern352 <f0><U+009F><U+0091><U+0080><f0><U+009F><U+0091><U+0080><f0><U+009F><U+0091><U+0080>1516646584 487458548 https://scontent-ort2-2.cdninstagram.com/vp/a7947eb9b112ae77765fc4de588978cb/5B17D374/t51.2885-19/s150x150/19761734_1414126255332842_7719441564023390208_a.jpgcyrbuiltsedan
17892821923152100 YGBSM 1516646811 31484972 https://scontent-ort2-2.cdninstagram.com/vp/082e02b38907eda2b7003c9a50cd7a73/5B06249A/t51.2885-19/s150x150/27577043_273972819803577_1611486401067483136_n.jpgisthisguyserious
17895336844131900 Beauty with the Beast #f16 my all time fav <f0><U+009F><U+009B><U+00A9><f0><U+009F><U+009B><U+00A9>1516646905 5880550922 https://scontent-ort2-2.cdninstagram.com/vp/e53f77cf21e1d98eaff8bc0b69ce56c2/5B0DAC55/t51.2885-19/s150x150/20987512_347130965723587_8001599182836072448_a.jpgsuman_ayush13
17919505357015800 Wow strong fight woman1516646962 2039822000 https://scontent-ort2-2.cdninstagram.com/vp/95e359675d5af205a6d9fec406545156/5B4E3182/t51.2885-19/s150x150/24838326_101231340667063_5819415055237644288_n.jpgahmedn22n
17918963671061900 Fire 1516646979 2039822000 https://scontent-ort2-2.cdninstagram.com/vp/95e359675d5af205a6d9fec406545156/5B4E3182/t51.2885-19/s150x150/24838326_101231340667063_5819415055237644288_n.jpgahmedn22n
17921023285040300 Too cool! @morgart_kayla1516647093 2017650474 https://scontent-ort2-2.cdninstagram.com/vp/1b1c915dada58ecd2c675f729fc1e19d/5B29F701/t51.2885-19/s150x150/24845327_135140523831942_6806056965404360704_n.jpgadrianna_morgart
17860535284208300 <f0><U+009F><U+0098><U+008D> in love !!!1516647293 191343376 https://scontent-ort2-2.cdninstagram.com/vp/a520c26468b0cb683d1f67e91e340ec1/5B1FD536/t51.2885-19/11910350_297402050430190_2026110713_a.jpgpepegrillo89
17907340705106600 @ihashubeita 1516647348 2345937556 https://scontent-ort2-2.cdninstagram.com/vp/d8ada6a718c57d9a59439dc73a3b2877/5B4BBC1D/t51.2885-19/s150x150/28151230_1729478153781410_7990125066035658752_n.jpgatilio_vurvopolos
17911435531078600 LOVE this!! 1516647427 27753143 https://scontent-ort2-2.cdninstagram.com/vp/64851f367870effc3007913a7f459a8e/5B202F43/t51.2885-19/s150x150/23347487_1707081442677369_7867464622582792192_n.jpgjacqgould
17920140301030800 A dream that i will never accomplish ...1516647437 3245473458 https://scontent-ort2-2.cdninstagram.com/vp/a9c38ec3cc104250afc258b250857b70/5B21E475/t51.2885-19/s150x150/27576504_165985677365135_2541958432584368128_n.jpgcall_me_numpt
17910967678076700 @charlie_collie 1516647440 15755357 https://scontent-ort2-2.cdninstagram.com/vp/8b37724a96371a4b90aeea9cf95bb7a5/5B04F0EB/t51.2885-19/s150x150/26265550_155658871750736_768762089006170112_n.jpgpassion4nyc
17920419412009400 Does she have Instagram? <f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D><f0><U+009F><U+0098><U+008D>1516647462 5527825192 https://scontent-ort2-2.cdninstagram.com/vp/77ddab8d0ad9e53ff61736bafb85fa75/5B11E2F4/t51.2885-19/s150x150/18723116_1696264410677425_7877078713601359872_n.jpgaraujo253
17905484125127500 Very good 1516647881 6796434290 https://scontent-ort2-2.cdninstagram.com/vp/bf1c9154fd05177b08c3a973cf4fd1bb/5B2AC2D4/t51.2885-19/s150x150/25018195_345706149227073_2352056333332643840_n.jpgaminallah.a
17889439036173600 #TeamShaw 1516647902 365607661 https://scontent-ort2-2.cdninstagram.com/vp/435c11c5ce19fdb69cf1d79c99b50ed6/5B055F3A/t51.2885-19/s150x150/14583454_679701112186705_6361606284961120256_a.jpgbeningold318
17861953780206500 @anna.ekstromh ålder är bara en siffra<f0><U+009F><U+0099><U+0084>1516648016 2222529160 https://scontent-ort2-2.cdninstagram.com/vp/9b6c447f0cb7df0c14a7f5df3fc71a8f/5B1D70E5/t51.2885-19/s150x150/27877706_204447570291007_9108200985573982208_n.jpgville360
17902468441091800 <f0><U+009F><U+0098><U+00A5><f0><U+009F><U+0092><U+00AA>1516648216 5910717791 https://scontent-ort2-2.cdninstagram.com/vp/aa780dcaeb1acedc96ad291f3a9296de/5B181D3A/t51.2885-19/s150x150/20986994_265826810572415_4917742609096507392_a.jpgshocktherapy.jn
17895479689186600 <f0><U+009F><U+0091><U+008D><f0><U+009F><U+008F><U+00BC><f0><U+009F><U+0087><U+00BA><f0><U+009F><U+0087><U+00B8>1516648412 1370345057 https://scontent-ort2-2.cdninstagram.com/vp/50ea7f8efbe5aaffd3e871d7967a27a5/5B1837AF/t51.2885-19/s150x150/11258372_117880065219772_1152994460_a.jpgjoeypr69
17907772753102700 BIUTIFUL 1516648554 2870647697 https://scontent-ort2-2.cdninstagram.com/vp/dcd3a183478cefc2d2d8dd68d43ce8cd/5B1ACB4B/t51.2885-19/s150x150/23735234_1946590205607281_4776352826878066688_n.jpgshahin_e_aseman
17893612492192200 @jonathan_viv123 I can't wait to see you paired up with someone like this. <f0><U+009F><U+0098><U+0087><f0><U+009F><U+00A4><U+0097><f0><U+009F><U+0098><U+0098><f0><U+009F><U+0087><U+00BA><f0><U+009F><U+0087><U+00B8><f0><U+009F><U+0087><U+00BA><f0><U+009F><U+0087><U+00B8>1516648563 1184245960 https://scontent-ort2-2.cdninstagram.com/vp/14fa7c884c0014626856e74f1e665da7/5B0C1754/t51.2885-19/s150x150/13151013_891647477630324_381289661_a.jpgjmonje14
62
Bibliography
Aggarwal, C. (2011). Social Network Data Analytics. Vasa. https://doi.org/10.1007/978-
1-4419-8462-3
Aggarwal, C. & Yu, P. S. (2005). Online Analysis of Community Evolution in Data
Streams. Proceedings of the 2005 SIAM International Conference on Data Mining,
56–67. https://doi.org/10.1137/1.9781611972757.6
Andriopoulou, F. & Lymberopoulos, D. K. (2012). Artificial Intelligence Applications
and Innovations. IFIP Advances in Information and Communication Technology
(Vol. 382). https://doi.org/10.1007/978-3-642-33412-2
Arcega, R. (2018). instagram-scraper. Retrieved from
https://github.com/rarcega/instagram-scraper
Ashraf, S. S., Verma, S., & Tech, M. (2016). A Survey on Sentiment Analysis
Techniques on Social Media Data, 3(3), 65–68. Retrieved from
http://epubs.siam.org/doi/abs/10.1137/1.9781611972757.6
Barberá, P. (2017). instaR. Retrieved from https://github.com/pablobarbera/instaR
Bian, J., Yoshigoe, K., Hicks, A., Yuan, J., He, Z., Xie, M., Guo, Y., Prosperi, M.,
Salloum, R., Modave, F. (2016). Mining Twitter to Assess the Public Perception of
the “Internet of Things”. PLoS ONE, 11(7), 1–14. Retrieved from
http://10.0.5.91/journal.pone.0158450
Chan, C. K., Vasardani, M., & Winter, S. (2014). Leveraging Twitter to Detect Event
Names Associated with a Place. Journal of Spatial Science, 59(1), 137–155.
Retrieved from http://10.0.4.56/14498596.2014.852073
Chaykowski, K. (2016). Instagram’s Big Picture. Forbes, 198(2), 62–69. Retrieved from
http://widgets.ebscohost.com/prod/customerspecific/ns000290/authentication/index.
php?url=http://search.ebscohost.com/login.aspx?direct=true&AuthType=ip,cookie,s
hib,uid&db=heh&AN=117050301&lang=pt-br&site=eds-
live&scope=site%5Cnhttp://content.ebscohost.com
Cheng, J. (2017). leaflet. Retrieved from https://cran.r-
project.org/web/packages/leaflet/leaflet.pdf
Erlandsson, F., Bródka, P., Borg, A., & Johnson, H. (2016). Finding Influential Users in
Social Media Using Association Rule Learning. Entropy, 18(5), 1–15. Retrieved
from http://10.0.13.62/e18050164
Glez-Peña, D., Lourenço, A., López-Fernández, H., Reboiro-Jato, M., & Fdez-Riverola,
63
F. (2013). Web Scraping Technologies in an API World. Briefings in
Bioinformatics, 15(5), 788–797. https://doi.org/10.1093/bib/bbt026
Hello World - GitHub Guides. (n.d.). Retrieved February 3, 2018, from
https://guides.github.com/activities/hello-world/
Injadat, M., Salo, F. & Nassif, A. B. (2016). Data Mining Techniques in Social Media: A
Survey. Neurocomputing, 214, 654–670. Retrieved from
http://10.0.3.248/j.neucom.2016.06.045
Instagram Scraper - Im Risto - Internet Marketing Blog. (n.d.). Retrieved February 13,
2018, from http://imristo.com/social-media-tools/instagram-scraper/
Jang, S. & Lee, E. (2009). Reliable Mobile Application Modeling Based on Open API. In
D. Śl\kezak, T. Kim, A. Kiumi, T. Jiang, J. Verner, & S. Abrahão (Eds.), Advances
in Software Engineering (pp. 168–175). Berlin, Heidelberg: Springer Berlin
Heidelberg.
JSON. (n.d.). Retrieved February 13, 2018, from https://www.json.org/
Leese, C. B. (2015). Mining Social Media for Intel. U.S. Naval Institute Proceedings,
141(8), 82–83. Retrieved from
http://search.ebscohost.com/login.aspx?direct=true&db=a9h&AN=109023176&site
=ehost-live
Li, H., Zha, Z.-J., Huet, B., & Tian, Q. (2016). Guest Editorial: Large-Scale Multimedia
Content Analysis on Social Media. Multimedia Tools and Applications, 75(3),
1365–1369. https://doi.org/10.1007/s11042-016-3255-z
Loftus, E. F. (2011). Intelligence Gathering Post-9/11. American Psychologist, 66(6),
532–541. Retrieved from http://10.0.4.13/a0024614
Lv, J. & Guo, J. (2016). Mining Communities in Social Network Based on Information
Diffusion. IEEJ Transactions on Electrical & Electronic Engineering, 11(5), 604–
617. Retrieved from http://10.0.3.234/tee.22278
Mezghani, M., Péninou, A., Zayani, C. A., Amous, I., & Sèdes, F. (2017). Producing
Relevant Interests from Social Networks by Mining Users’ Tagging Behaviour: A
First Step Towards Adapting Social Information. Data & Knowledge Engineering,
108, 15–29. Retrieved from http://10.0.3.248/j.datak.2016.12.003
Milgram, S. (1967). The Small-World Problem. Psychology Today.
https://doi.org/10.1007/BF02717530
64
Our Story – Instagram. (n.d.). Retrieved February 12, 2018, from https://instagram-
press.com/our-story/
Peng, R. (2015). The Reproducibility Crisis in Science: A Statistical Counterattack.
Significance, 12(3), 30–32. Retrieved from http://10.0.4.87/j.1740-
9713.2015.00827.x
Peng, R. D. (2009, July). Reproducible Research and Biostatistics. Biostatistics, pp. 405–
408. Retrieved from http://10.0.4.69/biostatistics/kxp014
Peng, R. D. (2011). Reproducible Research in Computational Science. Science,
334(6060), 1226–1227. Retrieved from
http://search.ebscohost.com/login.aspx?direct=true&db=eue&AN=69861275&site=
ehost-live
Pitic, A. G., Volovici, D., Tara, A., & Mite, A. C. (2013). Data Mining in Social
Networks. Revista Transilvania, (2), 32–35. Retrieved from
http://search.ebscohost.com/login.aspx?direct=true&db=a9h&AN=89005274&site=
ehost-live
Platform Changelog - Instagram Developer Documentation. (n.d.). Retrieved February
14, 2018, from https://www.instagram.com/developer/changelog/
Postaddict.me. (2018). instagram-php-scraper. Retrieved from
https://github.com/postaddictme/instagram-php-scraper
R: What is R? (n.d.). Retrieved February 13, 2018, from https://www.r-
project.org/about.html
Richelson, J. T. (2015). The US intelligence Community. Westview Press.
Ronczkowski, M. R. (2011). Terrorism and organized hate crime: Intelligence gathering,
analysis and investigations. CRC Press.
usairforce. (2018). U.S. Air Force (@usairforce) - Instagram Photos and Videos.
Retrieved February 18, 2018, from https://www.instagram.com/p/BeQuT-
8HfpC/?taken-by=usairforce
Ven, K. & Verelst, J. (2006). The Organizational Adoption of Open Source Server
Software by Belgian Organizations. In E. Damiani, B. Fitzgerald, W. Scacchi, M.
Scotto, & G. Succi (Eds.), Open Source Systems (pp. 111–122). Boston, MA:
Springer US.
Wickham, H. (2015). R Packages. Sebastopol, CA: O’Reilly Media. Retrieved from
http://r-pkgs.had.co.nz/
65
Wickham, H. & Grolemund, G. (2016). R for Data Science. Sebastopol, CA: O’Reilly
Media. Retrieved from http://r4ds.had.co.nz/
Wickham, H. & Henry, L. (2017). tidyr: Easily Tidy Data with 'spread()' and 'gather()'
Functions. R package version 0.7.2. https://CRAN.R-project.org/package=tidyr
66
REPORT DOCUMENTATION PAGE Form Approved OMB No. 074-0188
The public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of the collection of information, including suggestions for reducing this burden to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to an penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number.
PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS.
1. REPORT DATE (DD-MM-YYYY)
22-03-2018 2. REPORT TYPE
Master’s Thesis
3. DATES COVERED (From – To)
October 2016 – March 2018
TITLE AND SUBTITLE
An Open Source Approach to Social Media Data Gathering
5a. CONTRACT NUMBER
5b. GRANT NUMBER
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S)
Kallhoff, Anthony J., Second Lieutenant, USAF
5d. PROJECT NUMBER
5e. TASK NUMBER
5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAMES(S) AND ADDRESS(S)
Air Force Institute of Technology
Graduate School of Engineering and Management (AFIT/EN) 2950 Hobson Way, Building 640
WPAFB OH 45433-8865
8. PERFORMING ORGANIZATIONREPORT NUMBER
AFIT-ENS-MS-18-M-130
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
Intentionally Left Blank10. SPONSOR/MONITOR’SACRONYM(S)
11. SPONSOR/MONITOR’S REPORTNUMBER(S)
12. DISTRIBUTION/AVAILABILITY STATEMENTDISTRUBTION STATEMENT A. APPROVED FOR PUBLIC RELEASE; DISTRIBUTION UNLIMITED.
13. SUPPLEMENTARY NOTES
This material is declared a work of the U.S. Government and is not subject to copyrightprotection in the United States.
14. ABSTRACT
Modern usage of social media affords the military intelligence and analytic communities novel approaches to gather information. However, the tools and resources to develop these methodologies are still maturing. Furthermore, current data acquisition tools are not available to the DoD for all social media platforms. This thesis addresses a small subset of this problem by developing an open source methodological approach to collect and manage data from a popular social media site that has previously been inaccessible to defense intelligence organizations. This approach was operationalized via the R package called instaExtract, and an exemplar analysis was performed to demonstrate its application and efficiency for intelligence gathering.
15. SUBJECT TERMS
Data Mining, Social Media, Instagram, Data Acquisition, Data Tidying, Open-Source, R, GitHub,
Web-based Hosting 16. SECURITY CLASSIFICATION OF:
17. LIMITATION OF ABSTRACT
UU
18. NUMBER OF PAGES
76
19a. NAME OF RESPONSIBLE PERSON
Dr. Bradley Boehmke, AFIT/ENS a. REPORT
U
b. ABSTRACT
U
c. THIS PAGE
U
19b. TELEPHONE NUMBER (Include area code)
(937) 255-3636
([email protected])Standard Form 298 (Rev. 8-98) Prescribed by ANSI Std. Z39-18