Mapping Political Extremism on Twitter in Brazil

Abstract

Data from Social Media Networks may contain individual level information that is rarely available in institutional databases. It can reveal subjective and immaterial aspects that can help us to understand social behavior and, when geolocated, it may also enable understanding of spatial patterns and relationships. There is a wide range of studies that address the relationships between Twitter use and political campaigns and how it has been used in several countries to drive debates, news agenda and influence public opinion. This paper focuses on analyzing the relationship between Brazil’s 2022 election results, the most polarized in recent years, and Twitter geolocated data regarding the January 8th events in Brazil, when the invasion of the National Congress and the Supreme Court in Brasília took place. The goal was to understand what the territory might reveal about this relationship. The state of Paraná was selected because it is from where many demonstrators left for Brasilia. We performed visual exploratory data analysis and spatial data analysis at municipality scale. As a result, we can cite a strong relationship between the municipalities that have more votes in each of the candidates and the number of tweets in favor or against the far-right extremist activities, and that a significant number of individuals turned against the act, especially after the media showed the damage caused to public property.