Impacto da pandemia da COVID-19 nos padrões de crimes: análises espaço-temporais para avaliar o passado e informar o presente

Abstract

Social distancing is one of the most effective measures to reduce the spread of COVID-19. However, its practice changes people’s routines, impacting crime patterns. This work has two primary objectives: (i) to analyze the impacts of the pandemic on crime patterns in Curitiba and (ii) to implement a tool to identify and alert changes in crime patterns. Each of the crimes was analyzed individually. Analyzes were performed on the time series through the additive decomposition of its components to identify trends and changes in temporal patterns. Regression models were built to measure the influence of restriction measures and the population’s permanence at home on the pattern of crimes. Local indicators of spatial association were used to identify relevant spatial clusters in the entire region of Curitiba. The analysis identified several statistically significant crime pattern changes. With the knowledge acquired in the studies, a pattern change detection and alerting tool were proposed. The developed tool uses time series and geographic data anomaly detection to identify pattern changes. The tool can be used in any crime tracking scenario, not necessarily associated with pandemics.