what’s wrong with my pie chart?!accessola2.com/superconference2016/sessions/1200wha.pdfwhat’s...
Post on 17-May-2018
225 Views
Preview:
TRANSCRIPT
WHAT’S WRONG WITH MY PIE CHART?!
Jesse Carliner, Acting Communications Librarian & Reference Librarian Nicholas Worby, Government Information & Statistics Librarian
January 28, 2016
Developing a data visualization program from the ground up
OLA Super Conference 2016
WHAT TO EXPECT?
● HUGE field - check out #dataviz on Twitter● For information professionals ● NOT for designers or data viz pros ● Generalists -- add a new tool to your toolkit● Limited resources
OUR CONTEXT
● University of Toronto Libraries● Robarts Library - Humanities and Social Sciences Library● Reference & Research Services Department, including
Government Information and Statistics ● Map & Data Library
OUR CONTEXT Evolving User Needs
OUR CONTEXT
Increasing consultation requests
OUR CONTEXT
How did we get involved in data visualization?
WHY DATA VISUALIZATION?● Natural extension of what we already do● User demand ● Relevance● Data literacy● Form of scholarly communications● Professional practice● Another way to use our resources
WHY NOT?● Huge field● Many discipline-specific practices & requirements● Never going to be experts● Limited resources
DATA VIZ IN ACADEMIC LIBRARIES● What types of services are offered?● Which types of staff offer these services and what are
their professional/academic backgrounds?● Where are data visualization services located
administratively within institutions?● What types of software and hardware are being used?
SERVICES
● Instruction (e.g. software training, fundamentals of data visualization)
● Online learning objects (e.g. tutorials)● Visualization labs● Consultations● Drop-in help
STAFFING & ADMINISTRATION
DATALIBRARIANS
GISLIBRARIANS
VISUALIZATIONLIBRARIANS
MAPS, GIS OR DATA SERVICES
DIGITAL SCHOLARSHIP UNITS
VISUALIZATION ANALYSTS
PHD STUDENTS FACULTY
FACULTY RUN LABS WITHIN LIBRARIES
Visualization service providers
SOFTWARE
… plus other statistical software packages & GIS programs
HARDWARE & FACILITIES
NCSU Visualization LabDuke University Data Visualization Services Lab
University of RochesterUniversity of Michigan 3D Lab
● Presentation spaces● Immersive, 3D displays● Teaching spaces● High-powered
computer workstations & computing clusters
● It varies...
HOW WE GOT STARTEDInitial instruction request:● 50-minute session on best practices for data visualization
in scholarly publications including:○ an overview of useful tools○ a practical demonstration of critiquing and revising a
visualization, provided by the professor
IMMEDIATE CHALLENGES● Can we even scratch the surface in 50 minutes?● What are the norms and best practices for
visualization for the discipline and the data?● Which tools should we use?● How do we get up to speed with those tools?● How much professional development is too much
professional development?!
BUILDING CAPACITYBooks:
Edward Tufte. (1983). The Visual Display of Quantitative Information.
Katy Borner & David E. Polly. (2014). Visual Insights: A Practical Guide for Making Sense of Data.
Alberto Cairo. (2013). The Functional Art.
Colin Ware. (2004). Information Visualization: Perception for Design.
BUILDING CAPACITYFree or low-cost online courses:
Information Visualization MOOC
Instructor: Katy Borner
http://ivmooc.cns.iu.edu
Data Visualization and Infographics with D3.js
Instructor: Alberto Cairo Scott Murray
http://journalismcourses.org/D3.html
● Python● R● Cleaning data● & more!
www.coursera.org
BUILDING CAPACITY
Blogs & social media:
Nature: Points of View Twitter: #datavizhttp://bit.ly/1juf0zm
Martin Krzywinski: VizBi Conferences:http://mkweb.bcgsc.ca/ http://bit.ly/1SHdtJT
BUILDING CAPACITY
Tutorial series:
Tableau:
http://www.tableau.com/learn/training
Duke Data Visualization Services Recordings:http://bit.ly/1rSwac4
BUILDING CAPACITY
Staff Training:● Library staff as instruction guinea pigs● Health Science Information Consortium of Toronto’s
annual professional development event
DEVELOPING WORKSHOPSKey considerations: Audience● What type(s) of data do they work with?● Are there discipline-specific visualisation norms, practices,
or forms?● What are your audience’s baseline tech skills like? Can they
program? Which languages?
DEVELOPING WORKSHOPSKey considerations: Resources
● How much time do you have?○ How much can you realistically cover?○ What software/techniques can students really learn?
● What computing resources are available to you?● What sample datasets can you use?
○ How much clean-up/prep will be required?
DEVELOPING WORKSHOPSStructure of 2nd workshop:
● Overview of data visualization workflow model● Best practices for visualization: theory & practice● Visualization critiques● Software tutorial● Participants create their own visualizations
DESIGN WORKFLOWModel adapted from Katy Borner’s Visual Insights
Audience & Purpose
Select and Prepare Data
Select Visualization
Form
Select Visualization
Elements
Visualize Data
Share
DEVELOPING WORKSHOPSStructure of 2nd workshop:● Overview of data visualization workflow model● Best practices for visualization: theory & practice● Visualization critiques● Software tutorial● Participants create their own visualizations
BEST PRACTICES: THEORY
Source: Cairo, Alberto. The Functional Art, p.120
Graphical perceptual accuracy,Cleveland & McGill 1984
BEST PRACTICES
Use consistent axes when comparing two adjacent graphs!
DEVELOPING WORKSHOPSStructure of 2nd workshop:● Overview of data visualization workflow model● Best practices for visualization: theory & practice● Visualization critiques● Software tutorial● Participants create their own visualizations
CRITIQUING VISUALIZATIONS
Differing scales?
Exaggerating data?
Appropriate colour choice?
Source: The Fraser Institute: http://bit.ly/1nuqPg6
Problems!
SOFTWARE TUTORIALWhy Tableau:
● Requires no programming● Drag & drop functionality● Supports a wide range of visualization types● Tableau Public is FREE!
SOFTWARE TUTORIALApproach:
● Highly-structured, 45-minute tutorial creating basic visualizations using prepared datasets
● Used public health, economic, and humanities datasets○ Data selection & prep most time-consuming part
of preparation
ASSESSING OUR WORK
Well-attended
Positive feedback
ASSESSING OUR WORK
1. Were your overall expectations of the workshop met? [ 30 ] Yes [ 0 ] Somewhat [ 0 ] No
2. Did you find the pace of the workshop?[ 1 ] Too fast [ 29 ] Just Right [ 0 ] Too Slow
3. Did you find the concepts and theories useful?[ 26 ] Yes [ 4 ] Somewhat [ 0 ] No
4. Did you find the Tableau tutorial?
[ 2 ] Too basic [ 27 ] Just Right [ 1 ] Too Advanced
5. How likely are you to use what you learned today in your professional practice?[ 19 ] Definitely [ 11 ] Maybe [ 0 ] No
HSICT Workshop Feedback
ASSESSING OUR WORK
“Every aspect within a visualization has a purpose.”
“I will take greater care to ensure that the message I want to convey in my visualizations, is the message that I intended to deliver. “
“Effective communication is so important in the workplace and I will definitely take into account all the principles talked about today.”
“The audience will get more out of a visualization if it is tailored to the data type.”
Graduate Student Feedback
ASSESSING OUR WORK
“I think clarification of my purpose and audience is very important which might just save me a lot of trouble later.”
“[I] realize that it is important to be a bit more selective in this process [of colour selection]. It had never occurred to me to ensure that my visuals were accessible to colour-blind people.”
“I learned the importance of choosing every visualization element carefully...these elements can be optimized to best convey data, as well as avoided in situations where trends are not actually there that would be falsely conveyed.”
Graduate Student Feedback
ASSESSING OUR WORK
“Ease of use also makes Tableau a great resource. I can work efficiently and display my visualizations in a more diverse way than Excel does (or that I know how to do within Excel).”
“Tableau...makes data visualization much easier than other tools. To this point, I have been using Microsoft Excel, which is much more counter-intuitive and time consuming.”
“I hope to continue to explore Tableau in the near future and investigate its use in my upcoming poster and PowerPoint presentations.”
Graduate Student Feedback
LESSONS LEARNED
● Can’t do it all or do it on your own● Huge demand and lots of opportunities● Focus and draw on what you know● Don’t let them bring their own data● There is no one perfect software● There’s always more to learn
FUTURE PLANS● Enlist more people and their expertise ● More workshops● Offer staff training● Pilot consultations ● More software and software support ● Revamp the data viz research guide● Hire a grad student with programming and data viz skills
And, the dream….get a data visualization lab!
QUESTIONS
QUESTIONS?Nicholas Worby
Government Information & Statistics Librarian
nicholas.worby@utoronto.ca
Jesse Carliner
Acting Communications Librarian & Reference Librarian
jesse.carliner@utoronto.ca
University of Toronto Libraries
THANK YOU!
IMAGE CREDITSMarketing icon, created by Yamini Ahluwalia from the Noun Project
Questions icon, created by Vicons Design from the Noun Project
Sapling icon, created by Michelle Zamparo from the Noun Project
Sisyphus icon, created by Nikolay Necheuhin from the Noun Project
Thumbs-up icon, created Centis Menant from the Noun Project
top related