July 14, 2007

VISP Visualization 0.4

Filed under: Portfolio,Projects,VISP — admin @ 6:30 pm

VISP Visualization 0.4

I had to make some changes. First, the agent-based model for packing fact circles was computationally intensive. In other words, as networks with larger sets of nodes were loaded the visualization become slow or just simply stalled. Lada pointed me to this circle packing algorithm. Hat tip to Weixin Wang, Hui Wang, Guozhong Dai, and Hongan Wang for a very useful solution. Here are some other additions:

  • added url text pop-up
  • automatically scales network to fit screen
  • user can change the default size of the base node
  • changed the variation of node sizes to be based on area

You can see it here

VISP Visualization 0.4

copy and paste “data/test.xml” into the get dataset text field.

June 4, 2007

VISP Visualization 0.2

Filed under: Portfolio,Projects,VISP — admin @ 4:16 pm

Website Fact Visualization

This is the first draft of a visualization for the VISP project being conducted by Lada Adamic, Suresh Bhavnani, and Dale Hunscher. Each larger circle represents a webpage, the circles inside represent facts, and arrows represent hyperlinks. The intensity of the outer circle indicates the amount of traffic it receives from a search engine (i.e. darker is more traffic). The size of the inner circle represents the amount of space a page dedicates to a given fact.

There are two noteworthy things. First, while it only calls one data set currently, it is designed to be read in any properly formatted XML data set. That means that it could be extended to include any webpage and search queries. Second, you may notice that the circles seem to jiggle around. It’s not purely for aesthetic reasons, honest. Because any data set could have n number of facts at any given size, it is difficult to write an algorithm for all scenarios. It is much easier (as in this case) to simply treat facts as autonomous agents with a strategy, get as close to the center as possible. What we see are the fact circles trying to find more compressed configurations.

The visualization can be seen here. Be sure to copy and paste this path “data/melanoma_com.xml” to the data set in the text field and simply press the “get dataset” button.