Crime geo-surveillance in microscale urban environments: NetSurveillance

Events and phenomena such as crime incidents and outbreak of an epidemic tend to form concentration of high risks known as hotspots. Geosurveillance is an increasingly popular notion for detecting and monitoring the emergence of and changes in hotspots. Yet the existing range of methods are not designed to accurately detect emerging risks at the micro-scale of street-address level. This study proposes NetSurveillance, a method for monitoring the emergence of significant concentration of events along the intricate network of urban streets. Through a simulation test, the study demonstrates the high accuracy of NetSurveillance in detecting such clusters, and outperforms its conventional counterpart conclusively when applied at the individual street address level. Empirical analysis of drug incidents from Chicago also illustrates its capacity to identify rapid outburst of crimes as well as a more gradual build-up of such concentration, and their disappearance, either as a one-off or as part of a recurrent hotbed