Discovering Influential Members of Congress

In the US Congress, a legislator can endorse a bill publicly before the vote to determine whether it is passed on to the president to sign by co-sponsoring it. Although there is no limit at this time on number of co-sponsors, a legislator only co-sponsors 2-3% of all the bills. Thus, the legislators make considerable effort in deciding which bills to co-sponsor and the network of co-sponsorship can lead us to interesting insights into the machinations of American federal politics. Congressional co-sponsorship is a driver of bill passage in the US Congress, but attempts to understand how co-sponsorship affects bill passage and which legislators most directly influence bill passage are still in their infancy. For our course project, we study the network structure of legislators by considering the co-sponsorships as interaction edges amongst them. First, we reproduce results from previous researchers and report basic measures of network connectivity of nodes, and link analysis results such as PageRank and HITS which have never been tried on this dataset. Secondly, we attempt to derive a more effective predictor of legislative success than has been found thus far. The study of influence in networks has advanced a great deal in the last few years through research on viral marketing. Measures of influence and cascades are applied to this dataset to find new insights into US politics.