Hotbeds of crime and the search for spatial accuracy

One of the most important aspects of spatial crime analysis is the identification of hotspots: areas of the highest crime concentration. This paper advances a methodology for hotspot detection based on a global moving window approach combined with the use of local statistics to define the hotspot limit. This technique generates hotspots that both follow the urban morphology of the crime distribution and ensures their spatial segregation. The hypothesis that police officers can construct an accurate perception of crime distribution from exposure to daily policing practices is used to demonstrate an application in the use of hotspot analysis. Significant regions generated from recorded crime data are compared with perceived local hotspots catalogued from surveys with police officers. Results from this study show two discrete types of hotspot, here termed hotpoints and hotbeds. The morphology of these crime hotpoints and hotbeds is discussed and possible causes documented. Received: 9 October 1998/Accepted: 22 September 1999