Community Detection within the US House based on Policy Voting

With an increase in the rise of divisive politics that are often contrarian to the core beliefs and manifesto of political parties, it is become imperative for political parties and candidates to find a solid middle ground where they both compromise. However, this compromise cannot be integrated at time of voting which is why it is very important to cluster defectors. Clustering analysis is instrumental in the process of understanding possible relationships between defectors and their original party. For instance, if a defector consistently votes against his own, they simply wouldn't be part of it but that is not the case. Defections usually happen due to a specific policy or more importantly, the domain of a political policy. As such, seeing where the defector concurs with the opposition is desired objective of the proposed study. The proposed model utilizes the Louvain Walk as aclustering methodology due to its focus on modularity measures to understand the vitality of a node to a particular subnetwork and cluster while also putting an emphasis on the cardinality of its relationship with those amongst its cluster. Due to existing target labels, it'll also be uncovered what policies and which states cause the most defections and how they interact with each other which can only be done bygraph analysis.