A street corner-level methodology to analyze the influence of points of interest on urban crime

As cities have evolved, so too have crimes, becoming increasingly sophisticated, violent, and intense. This evolution has pushed security models to their breaking point, rendering many traditional strategies obsolete in the face of these new challenges. Consequently, society, especially law enforcement agencies, needs more sophisticated tools to assist them in decision-making. The growing digitization of data over the last decade has enabled the large-scale and highly agile collection of urban data which can be exploited to conduct crime analysis tasks and in particular to identify relevant crime patterns. In this study, we present a computational methodology to investigate the relationship between crime occurrences and the proximity to points of interest (POIs) within a city. In particular, this methodology can perform a segmented analysis, according to socioeconomic patterns of different city regions, using clustering algorithms. Through case studies in the Brazilian cities of Maceió and Arapiraca, we validate the proposed methodology and demonstrate a global correlation between POIs and crime occurrences in both cities. Furthermore, this correlation varies significantly when analyzing street corners segmented by socioeconomic patterns and across both cities. These findings validate the proposed methodology and demonstrate that this approach provides a robust framework for strategic decision-making, enabling law enforcement agencies to allocate resources more effectively and enhance overall public safety.