<-- Advertise here.

The Clear Congress Project [clearcongressproject.com] by Thomas Gibes is a real-time visualization of US Congress data made available through the Sunlight Labs' Real Time Congress API, Google News, Twitter, and other data sources. The aim of the project is to serve as a possible model for facilitating governmental transparency beyond simple data access by proposing a new format for polical news distribution.

The scatter plot shows the Representatives and Senators as a collection of two-colored bubbles that are mapped against 2 axes: Leadership (i.e. sponsoring popular legislation versus other's legislation) and Partisanship (i.e. Democratic versus Republican). The size of each bubble denotes the number of bills that were sponsored by the according politician, while the connecting lines reveal the sponsor/cosponsor relationships.

The Leadership score was calculated by subtracting the number of times a legislator cosponsors another legislator's bills from the number of cosponsors that a legislator's sponsored bills attracts. The Partisanship score was derived by the number of times a legislator works with members of their own party, placing them further to the left of the spectrum the more they work with Democratic members and further to the right of the spectrum the more they work with Republicans.

From this graph, one could for instance discover how all the members of the Tea Party Caucus show up above the mid-leadership line and on the extreme right (Republican side), or how the two current legislative leaders, Eric Cantor and Jim Gerlach, have garnered large amounts of historical bipartisan support for their past legislation. Alternatively, one can also perceive how the legislators who just signed a furious letter to Obama are all placed on the left side of the diagram, have no strong legislative leadership records, though still make up all of the Democratic leaders in Congress.




interesting: but the same old problem with all dot connected by lines charts -> lines represent links (or relationships) between dots, but the lenght of each line is decided only for the need of graphic representation and does never contain information: having different lenghts or different coordinates for each dot causes us to belive that some links are more important than others, while they should be shorter only when they represent stronger relationships (closer dots). The entire map should not be mapped on actual coordinates like "how much republican acts you did" instead dots should be placed closer to dots they "speak" a lot, and far from dots they never had direct relationships.

Wed 13 Jul 2011 at 7:35 PM

@simo: Hmm, I understand your point here, but one of my goals was to combine a scatterplot and network graph on the same plane, which means I could not encode information along the length of the line without messing with the formation created by the scatterplot. I would like to point out, however, that a small amount of information IS encoded on the connecting lines - the bolder the line, the more times that person has co-sponsored the highlighted persons legislation (there's a slight alpha on the line so the more times a connection is drawn, the bolder the line gets - it's not as visible later in a congressional session when more legislation is passed, something I will have to think about addressing, perhaps by encoding this information on the width of the line).

There is similar work to what you've described, using force-directed network graphs. Check out Schneiderman's SocialAction, which was one inspiration for this project: http://www.cs.umd.edu/hcil/socialaction/

Thu 14 Jul 2011 at 4:26 AM

Thanks Thomas for your reply and for considering my point. Using the lenght of a connection line as a way to display information was just an idea that came trying to make sense of what i was seeing in your chart, i didn't know it was actually used by someone, your link @ umd.edu is also interesting, but i cant find on their site this precise feature, i'll look deeper, thanks man :-)

Mon 25 Jul 2011 at 3:57 PM
Commenting has been temporarily disabled.