Brazilian Congress structural balance analysis

TL;DR

This paper analyzes the voting behavior of Brazilian politicians from 2011 to 2016 using signed social networks and clustering algorithms to understand political loyalty, coalitions, and polarization.

Contribution

It introduces a novel application of correlation clustering to Brazilian Congress voting data, revealing insights into political dynamics and social phenomena.

Findings

Identification of political coalitions and polarization patterns

Insights into loyalty and leadership within Congress

Detection of social phenomena like mediation and crisis

Abstract

In this work, we study the behavior of Brazilian politicians and political parties with the help of clustering algorithms for signed social networks. For this purpose, we extract and analyze a collection of signed networks representing voting sessions of the lower house of Brazilian National Congress. We process all available voting data for the period between 2011 and 2016, by considering voting similarities between members of the Congress to define weighted signed links. The solutions obtained by solving Correlation Clustering (CC) problems are the basis for investigating deputies voting networks as well as questions about loyalty, leadership, coalitions, political crisis, and social phenomena such as mediation and polarization.