Spatial analysis techniques applied to local crime patterns

Abstract

Despite growing crime rates, and increased computerisation of crime data within the police force, surprisingly little attention has been payed to techniques of analysis that could be applied to this data. This thesis investigates spatial analytical and statistical techniques which may be used for this task, and proposes a Bayesian forecasting technique, allowing the objective pattern detection mechanisms supplied by the quantitative examination of past data to be combined with the subjective knowledge of police officers. This method, along with others, is incorporated into a software package which may be run at a police station (Subdivision). Finally the software package is evaluated by members of the police force.