Information Mining and Visualization of Data from the Brazilian Supreme Court (STF): A Case Study

This paper describes a joint research of the Law School (Direito Rio) and the Applied Math School (EMAp) of the Getulio Vargas Foundation (FGV), Brazil to analyze information from judicial activities in some of the Brazilian courts. The data for the study included the entire collection of judicial decisions from 1988 to the present. The idea was to identify bottlenecks in the judicial processes at the STF. Introduction Large collections of textual data present a substantial challenge for extraction of relevant bits of information to feed subsequent statistical analysis and visualization pipelines. The peculiarities of the knowledge domain often require the implementation of customized natural language processing pipelines, along with specific knowledge organization systems, to describe the relevant terminology. This paper describes a joint research made by the Law School (Direito Rio) and the Applied Math School (EMAp) of the Getulio Vargas Foundation (FGV), Brazil. After initial contacts a joint venture was established between researchers from EMAp and Direito Rio to analyze the information from judicial activities, in some of the Brazilian courts. Initially, the Law School intended to analyze the behavior of the Brazilian Supreme Court (STF) to support public policy-making, and to identify bottlenecks in the judicial processes at the STF. The task was to analyze the texts of the entire set of recorded judicial decisions, accessible through the STF institutional site. This data had never been analyzed on this scale before, so a great deal of exploratory analyses was expected in order to reveal hidden patterns in the data. A number of a priori questions were proposed, basically aimed at determining if the STF was performing according to its constitutionally defined role, and if not, in what way it could be changed to better serve its purpose. Some of the methodology described herein was applied to the generation of the results published in the project‟s first technical report (“I Relatório – abril/2011 – O Múltiplo Supremo”, 2011). The results presented here go more in the general direction of exploratory data analysis.