A Spatio-Temporal Crime Prediction Model for Crime Prevention

Abstract

Development of effective crime prevention strategies requires understanding of the spatial/spatio-temporal occurrences of crime events and then predicting the number, place and time of crime events. In this respect, Geographical Information Systems (GIS) provide efficient tools for analyzing the spatial and spatio-temporal crime events. Similarly developing prediction models for criminal and terrorist activities can directly support crime prevention and counter-terrorism planning, respectively. This study presents a spatio-temporal crime prediction model generated by using time series forecasting with a simple spatial disaggregation approach in GIS. The proposed model is implemented for Bahcelievler and Cankaya Merkez Police Precincts' 2003 crime data obtained from Ankara Police Directorate. Although the proposed crime prediction model is developed for urban crimes, it can be adapted for terrorist activities by using appropriate scales in space and time.