polimaps: Supporting Predictive Policing with Visual Analytics

Abstract

Recently, predictive policing has gained a lot of attention, as the benefits, e.g., better crime prevention or an optimized resource planning are essential goals for law enforcement agencies. Commercial predictive policing systems commonly visualize predictions on maps but provide only little support for human analysts in the technical and methodological processes that constitute corresponding implementations. In this paper, we report on a project of bringing visual analytics to the field of predictive policing. We introduce a process model that includes machine learning as well as visualization and has been developed together with experts from a law enforcement agency. We also showcase a visual analytics tool, called polimaps, that is part of a real-world predictive policing project and implements elements of the proposed process.