Writing Code to Visualize Networks

From DevSummit
Jump to: navigation, search

How to visualize relational data - a (great) talk by Skye Bender-deMoll

Social graphs -- nodes and edges -- store the relationships between them

Explosion of tools for visualizing social graphs.

Can extract non-obvious information from looking at how people connect. -- Example, can figure out how people cluster into groups, social structures -- See how your base is connected and clustered and segmented

Great examples:

Touchgraph.net - TouchGraph photos

Sawmill strike example -- Figured out who were the nodes -- who you need to talk to convince the others. -- but really easy to distort the data -- we don't have good visual literacy around data visualization

Revealing Economic Terrorists: a Slumlord Conspiracy http://www.orgnet.com/slumlords.html

[Another time, talk about data cleaning tools, data normalization tools]

You can pull link organizational data out of Facebook via API.

Otherwise collecting data is very hard and expensive

Skye walked us through the current state of tools, referencing http://skyeome.net/projects/networkChart/comparison.html

Then he showed examples of data visualization used to tell stories.

Good examples

Prop 23 dirty oil - http://prop23.dirtyenergymoney.com/

Conservative foundations styrotopia.net/~unfluence/rightwing or http://angelsoftheright.net/

Then Skye discussed an open source project he has just started for embedding interactive network diagrams on a web page. nodeviz on code.google.com/p/nodeviz

Then we had a good discussion on Data Visualization. At the end of the day, still requires humans to make sense of the data, figure out what attributes mean, which are most important. Sometimes you want to tell the story, sometimes you want to just let people explore the data in a guided way, sometime you want to tell a story and allow people to verify the data underneath the story. Data visualization vs. data exploration vs guided tour of the data -- and being able to show the data.