| 
  • If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • Want to get organized in 2022? Let Dokkio put your cloud files (Drive, Dropbox, and Slack and Gmail attachments) and documents (Google Docs, Sheets, and Notion) in order. Try Dokkio (from the makers of PBworks) for free. Available on the web, Mac, and Windows.

View
 

L324 - Session 3

Page history last edited by David Zwarg 11 years, 10 months ago

Data visualization

 

Bridging the Gap (presentation)

w/Irene Ross and Yanick Assogba (IBM Social Media Research)

 

Participaroty & Civic Minded Online Vis.

http://sense.us

- used to explore cenus data

- cool: take snapshot of data & attach comments -- others come back & see your data + comments

 

http://many-eyes.com

- upload your own data

- visualize it

- discuss & comment like sense.us & embed

 

In many-eyes:

Treemap of Stimulus Bill

Put together by pro-publica.

Put together very shortly after the bill was published/proposed.

 

Examples:

http://crimemaps.com

(Stamen Design)

http://www.walkscore.com

http://www.walkshed.com

 

British Columbia Budget Vis.

http://blog.blprnt.com/blog/blprnt/bc-budget-visualization

Loss & Gain- done by an artist who was concerned w/arts funding

 

Newspaper Club

http://blog.newspaperclub.co.uk/2009/10/16/data-gov-uk-newspaper/

 

This We Know

http://www.thisweknow.org (GreenRiver.org, Intellidimension, Sway Design)

Links back to data where they got it from

Text can be really effective

 

Text Visualization

Treemap -- Newsmap: popular headlines organized in a treemap

Wordcloud/Wordle

Tag cloud -- originally built as a toy

Developer added a lot of work into the typography; aesthetics matter

 

Document Contrast Diagrams, Jeff Clark

What about larger document corpora?

14,000 documents?

 

Introducing IBM Visual Bill Explorer

"Collections" created by users

Break each bill into sections, each section has a topic

Color-coded by topic

Open to Beta testers

 

q: algorithmic categorization

q: what is the algorithm: machine learning based on historical legal documents

 

Tools:

http://many-eyes.com

http://swivel.com

http://www.tableausoftware.com/public

http://tables.googlelabs.com/

 

Protovis, Raphael, Processing.js, Prefuse

Processing, Java, Flash, Flare

 

http://github.com/sunlightlabs/clearmaps/

 

public google doc: http://bit.ly/gov20vis

 

Data Sources

data.gov

thomas.loc.gov

govtrack.us

www.regulations.gov

opencongress.org

 

public google doc: http://bit.ly/gov20vis

 

Discussion -- end of the spiel

 

What site do you use for word clouds?  What are the options?

http://www.many-eyes.com/

http://wordle.net/

 

How do you manage with different data formats? Do you use a framework?

-- we just use the stuff we can find in xml

-- example: house health care bill in xml, not senate health care bill

DTD for above xml documents is insane. Nothing special; 1 off implementations

Govtrack uses xml and RDF (converts automatically)

 

Implications -- you can lie well w/data you can lie better with visualizations.  What about the ethics of vis?

Visualizations are personal -- comments are used to track differences in opinions.  Comments linked to visualizations provide context.

 

Any guidelines for 'how to' interpret visualizations?

There are a couple documents that explain /what/ the visualizations are, but not necessarily /how to/ interpret them.

 

q1: What is many-eyes' mission?

q2: Are you relying on SEC documents? Enormous data store and enormous public service.  XMBL - standardize electronic filing.As part of IBM research - we do it because we are interested in 'text'.  It is interesting to visualize, and many-eyes is a way to parse information.

* Discussing them is an important part of many-eyes

* Be able to share URLs that represent every part of the visualization -- share the URL, you share the whole vis.

 

Visualizations are like opinions -- just another medium.

If more people use visualizations, they will be more educated, more informed, and know the difference between 100% in a pie chart.

 

Take data, load it into a tool.  Have you done work w/making many-eyes in order to work in another CMS, like Drupal.

1) many-eyes enables embedding.  Put a live version anywhere.

2) many-eyes wikified -- takes a feed and/or google docs spreadsheet

 

Basic tools:

Use Javascript -- browsers are getting better.

For a beginner: how simple is it to get started?

You need to be able to write the code.

 

MBTA - put out realtime data & developers started working with it. As developer, it was picked up and worked with.  Following the model from the MBTA, based on Chris Dempsey's talk, it seems like a successful model.

 

q: What should we visualizing?  What would be valuable?

 

What about non-traditional mapping?  Geographic tool kits.

TED talk from Hans roseling have you seen it?  Have you used time in your visualizations?

 

Does anyone do anything for mapping officials/their terms/etc.?  Have you worked with that data?

datamasher -- won Apps For America 2.0: pick 2 datasets, pick an operator & visualize it on a map

 

Interesting visualizations are personal maps.  The maps that individuals create to represent their environment.  Based on heart-rate, etc.

 

Chloropleth maps are becoming more popular in recent years. Also, new maps based on projections or other subjective factors are coming into the zeitgeist.

 

Maybe using audible synthesis -- not necessarily visualization, but making it accessible.

 

Really interesting problem: how do you make data accessible.  What do you do to make a visualization to the visually impaired.  How do you represent that data?  There is a spectrum of visual impairment, what would be compelling across the spectrum?

 

How to improve visualization reputations?  Would rating systems make them more compelling/credible? IBM doesn't have resources to verify authenticity of contributors (on many-eyes, etc).

 

If you are testing visualizations, what about using eye-tracking to provide feedback for quality and effectiveness?

 

Any insights as to what makes for an engaging vis?

 

Comments (0)

You don't have permission to comment on this page.