Uma análise exploratória de dados espaciais para criminalidade violenta no estado de Pernambuco

Abstract

This study aims to identify spatial patterns in violent crime rates in the state of Pernambuco, northeastern Brazil, based on the Economic Theory of Crime. Through the application of the rates of CVLI (Lethal and Intentional Violent Crime) and CVP (Violent Crime against Property), which represent the rates of homicide and robbery, in the period from 2014 to 2020, as variables of interest. For this proposal, the AEDE (Spatial Data Exploration Analysis) methodology was used, with the intention of testing spatial autocorrelation and identifying spatial clusters. With this, spatial clusters were identified, for both variables, in the mesoregions: Sertão, Agreste and RMR (Metropolitan Region of Recife). In addition, it was verified the rupture of the pattern of behavior in the year 2020 in different ways for each variable under study, as a possible effect of the covid-19 pandemic. It was also noted a possibility of future study regarding the increment of the income variable with the GDP (Gross Domestic Product) per capita of each municipality, by performing a spatial analysis with bivariate Local Moran's I, which resulted in low autocorrelation.