Clusterização aplicada à análise de dados da segurança pública: um estudo de caso aplicado no gerenciamento de recursos da Polícia Civil do Estado do Rio Grande do Norte (PCRN)

Abstract

In light of the emerging crime scenario in Brazilian society, public security agencies are increasingly concerned with the efficient allocation of material, technological, logistical, military, and, most importantly, human resources. In this context, this study proposes an evaluation of how the application of clustering techniques and data analysis can contribute to improving the efficient management of resources and the effectiveness of PCRN operations. Furthermore, it explores how these results can be replicated across other public security institutions in the country. The procedures that structure this work involve data collection and preparation, filtering of groups, and the removal of outliers from the variables using the z-score test, clustering based on similarity, and applying the silhouette metric. From this, it becomes possible to form clusters based on their characteristics, patterns, and relevant trends. The analysis of the results revealed the formation of four main groups among the studied units, which are categorized as follows: “Low investment in resources,” “focus on technology and weaponry,” “focus on human resources,” and “focus on mobility and logistics.” These groups exhibit similar characteristics, highlighting structural, physical, technological, and human factors that they share. Understanding these factors enables strategic decision-making by the organization’s upper management, aiming to achieve the best balance between utilized resources and demand satisfaction, thus ensuring the principle of efficiency in public administration. Additionally, based on the applied methodology, it will be possible to create standards and models that expand studies and applications for other public security agencies in the country.