visualizations, mashups and dashboards university of illinois at urbana-champaign

27
Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Upload: marilyn-tucker

Post on 18-Jan-2016

215 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Visualizations, Mashups and Dashboards

University of Illinois at Urbana-Champaign

Page 2: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Outline

• Visualization

• SEASR as a service

• Text App: Airport Data

• Text App: Correlation & Ngram Viewer

• Text App: ProseVis

• Attendee Project Work

Page 3: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Visualization Components

• JavaScript – GIS: GoogleMaps– Temporal: Simile– InfoVis: Protovis – Parallel Coordinates, Link Node, Arcs– InfoVis: D3 – Line Graphs, Tag Cloud

• GWT – Dendogram– Table Viewer

• Flash – InfoVis: Flare

• Applets– Data Mining Results: Decision Tree, Naïve Bayes, Rule

Association

• HTML– Reports

Page 4: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Mashups & DashboardsHTML driven dashboard• Connect to DB to

make the query• 6 analytical

services (flows)– 4 different tag

cloud views– 2 entity

relationship views

Example: SEASR, Protovis

Page 5: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Usage of Service from HTML

• Simple

<iframe src=http://meandre-dev.leo.ncsa.edu:10000/service/post?query=myquery width="100%" height="250”/>

Page 6: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

More Complex

<script>var servicePorts = [ 10000, 10001, 10002, 10003, 10004, 10005 ];var iframes = ["frame1", "frame2", "frame3", "frame4", "frame5", "frame6”];function searchService() {

var searchterm = document.getElementById("searchterm").value;

for (var i in servicePorts)window.frames[iframes[i]].location.href =

"http://meandre-dev.leo.ncsa.edu:" + servicePorts[i] + "/service/post?query=" + searchterm;}</script>…<iframe src="about:blank" name="frame1" id="frame1" width="100%" height="250”/>

Page 7: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Correlation-Ngram Viewer

Page 8: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Correlation-Ngram Viewer

• Pearson Correlation Algorithm

• 5 Flows run as ZigZag flows

Page 9: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Correlation-Ngram Viewer

• Words whose frequencies correlate with each other in a large dataset often turn out to have contextual associations as well

• Corrected version of the Google ngrams dataset (the English corpus)

• http://leovip029.ncsa.uiuc.edu/Correlation/

Page 10: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Correlation-Ngram Viewer

• New version of the Google ngrams viewer (for 1 grams)– addresses case-sensitivity

– period spellings

• Normalized to British practice, and period spellings have been normalized to modern practice (“honor” -> “honour,” “enow” -> “enough,” “publick” -> “public”)

– past-tense syncope (' d)

• based on 3-gram dataset to correct the treatment of syncopated past tenses (occurrences of “resolv’d” are now folded into the occurrences of “resolved”)

– f/s substitution as well as other OCR issues

• Searches within already stored correlation results (using Pearson) results for top 10K ngrams

• Computes correlation (using Pearson) results for given word against top 1K ngrams

Page 11: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Correlation-Ngram Viewer Flows

• Service Search Pearson for Top 10k Google Ngrams

– Uses the web service components to accept a query and executes a SQL query with this term against the table that stores the calculated Pearson results.

• Service Compute Pearson for Top 10k Google Ngrams

– Uses the web service components to accept a query and executes a SQL query to calculate the correlation of this search term against the top 10,000 ngrams.

• Service Query Pearson Results

– Shows the set of ngrams that have the highest Pearson score within the 10,000 result set.

• Service Google Ngrams Multi Line Chart

– Uses the web service components to accept a query, execute the SQL query with this term(s) to retrieve the frequency data and create a visualization showing the frequency graph for each ngram. The graph has a smoothing moving average that can be user adjusted.

• Service Google Ngrams List

– Uses the web service components to accept a query, execute the SQL query to identify the term(s) included in the graph and their total frequency counts. The list displays normalization for case, OCR errors from the transformation rules, and period spellings normalized to British.

Page 12: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Service Google Ngrams Multi Line Chart

Page 13: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Service Query

• http://leovip029.ncsa.uiuc.edu:10013/service/post?query=“dark”&query=“bright”&min_year=1700&max_year=1899

Page 14: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

HTML code calling Meandre Service

<script>

var servicePorts = [10012,10013,10014,10015];

var iframes = [ "frame0","frame1","frame2","frame3" ];

function searchService() {

var searchterm = document.getElementById("searchterm").value;

searchterm = searchterm.replace(/,,/g, ","); //remove extra commas

searchterm = searchterm.replace(/ /g, "");

searchterm = searchterm.toLowerCase();

var gsearchterm = searchterm.replace(/,/g, "%22&query=%22");

searchterm = searchterm.replace(/,/g, "%22,%22");

window.frames[iframes[0]].location.href = "http://leovip029.ncsa.uiuc.edu:" + servicePorts[0] + "/service/post?query=%22" + searchterm + "%22&min_year=1700&max_year=1899";

window.frames[iframes[1]].location.href = "http://leovip029.ncsa.uiuc.edu:" + servicePorts[1] + "/service/post?query=%22" + gsearchterm + "%22&min_year=1700&max_year=1899";

}

</script>

Page 15: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

ProseVis

Page 16: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

ProseVis• Visualization tool

• Seeks to identify other features than the "word" to analyze texts

• Features comprise sound – parts-of-speech– accent– phoneme– stress– tone– break index

• http://tclement.ischool.utexas.edu/ProseVis/

Page 17: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

ProseVis: Output Data

Page 18: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

ProseVis: Data Processing Flow

Page 19: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

ProseVis: Prosody Features

“Matisse” “Picasso”

Page 20: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Prosody Comparisons

Page 21: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

ProseVis: Prosody Comparisons

Tender Buttons

Page 22: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

ProseVis

• Calls Meandre service for data processing– Identify Prosody Features of Document

– Compare Prosody Features of Multiple Documents

Page 23: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Other Noteworthy Data Cleaning Tools• Data Wrangler

– http://vis.stanford.edu/wrangler/

– “Wrangler is an interactive tool for data cleaning and transformation.Spend less time formatting and more time analyzing your data.”

• Google Refine

– http://code.google.com/p/google-refine/

– “is a power tool for working with messy data, cleaning it up, transforming it from one format into another”

Page 24: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Demonstration

• Visualizations

• Text App: Correlation & Ngram Viewer

• Text App: ProseVis

Page 25: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Attendee Project Plan

• Study/Project Title

• Team Members and their Affiliation

• Procedural Outline of Study/Project– Research Question/Purpose of Study

– Data Sources

– Analysis Tools

• Activity Timeline or Milestones

• Report or Project Outcome(s)

• Ideas on what your team needs from SEASR staff to help you achieve your goal.

Page 26: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Learning Exercises

• Create a web page that accesses SEASR services

– Customize an HTML template that we have constructed to leverage 2-3 flows that you develop and execute

Page 27: Visualizations, Mashups and Dashboards University of Illinois at Urbana-Champaign

Discussion Questions

• What types of mashups or dashboards would you like to see?