Data driven smart policing: A novel road distance-based k-median model for optimal substation placement

Abstract

In the context of smart city research, finding patterns in crime data to explore trends in crime and to locate the presence of crime has been an exciting research field. Our aim in this paper is optimizing the location of police substations within the jurisdiction of a police station so that law enforcement agencies could work efficiently. These substations are placed near by road in order to immediately respond to nearby crimes. We attempt to optimize the average minimum distance to a crime from its nearest substation. This distance is found by the underlying road network. We are locating these substations in an infinitely large set of points in the given region. Case study has been done on real-world crime data from Lahore, Pakistan to show the efficiency of these methods. We also explored what are trends in crime data with respect to years, seasons and day/night and where to place substations accordingly. We evaluated the placement of substation by our model through comparing the average time taken to report to a crime against random substations placed within the same area. Last but not the least, the analytical insights provided by this study be useful for policy making and smart city research for the inclusion of sustainable cities and communities.