Data Mining the US Senate Votes

AJ writes "We used some old and new data mining techniques to see what was happening in the US Senate in 2003. Among other things, we identified the 'social' network of similarities between senators, how influential is each senator and each state, and a 3D VRML view of the Senate. You will be able to check whether Senator Kerry was a centrist or a liberal, and who is acting more cohesively, the Democrats or the Republicans. We provide our data and the source of all our analysis software (Orange and MPCA, both under GPL)."

Comments

Re:The Real Kerry

[Khomar]

Umm, they did this over the period he was FREAKING RUNNING FOR PRESIDENT.

Not true. Check the site again.

Gallery of the US Senate in 2003

This is data collected from last year not this year. Granted, he may have been gearing up for this, but has he been campaigning full time for two full years?

Re:interesting...

[elmegil]

You'll also note that they were among the four who had the highest non vote percentages, which is why they ended up clustered toward the majority. This is a side effect of running for president, not due to actual choices they would have otherwise made. Which means, they don't (necessarily) agree more with the Republicans at all; I think you'd want to look at a non-election season like 2001 or 2002, or else a longer stretch of the data to make that kind of determination.