Representação, visualização e análise de grandes volumes de dados urbanos espaço-temporais de segurança pública

Abstract

Technological transformations arising from the digital transformation of public bodies and the internet of things have enabled the generation and collection of large volumes of data, which can later be used in precise analyses. For public security, the application of new technologies in criminal analysis, patrolling and crime repression activities also bring countless possibilities, such as being able to visualize in depth the performance of the corporation’s actions, monitor criminal activity, understand patterns and seek alternatives to implement better security policies. Big Data is also no longer a distant reality for public security corporations. Tracking devices, surveillance cameras, monitoring systems, customer service systems and many other sources of information already provide a large volume of data that needs to be properly processed in order to obtain relevant knowledge. In recent years, many scientific works have proposed the use of Machine Learning algorithms to recognize spatial and temporal patterns of crimes. In this context, we propose in this work a visual analysis tool for spatio-temporal urban data, and a new algorithm for detecting crime spots. These solutions were validated using databases from the Military Police of the State of Alagoas (PMAL) in case studies where the objective was to analyze spatio-temporal data on crimes and patrolling. The results of this research will be important not only from a scientific point of view, but can also be used by PMAL to optimize its decision-making processes. A difference of this work is the approach of crime and patrol data sources in the same visualization tool.