Crime Mapping and Analysis in the Dansoman Police Subdivision, Accra, Ghana - A Geographic Information Systems Approach

Abstract

This work used Geographic Information Systems to create crime maps and analyze crime distribution within the Dansoman Police Subdivision in Accra, Ghana. It was done on a pilot basis and is intended to enable senior officers have a visual framework for prescribing action. Spatial data of 142 crime incidents namely Assault, Causing Damage, Defilement, Fraud, Rape, Robbery, Stealing and Unlawful Entry was analyzed using ArcGis®v10 software. A summary of the attribute table showed that Assault, Causing Damage and Unlawful Entry had the highest counts of 22 each while Rape and Stealing had the least counts of 12 each. Pin maps produced gave a graphical representation of crime locations against a backdrop of roads, existing police stations and district boundaries. Kernel density estimation maps showed the Mamprobi district had a high density of crime even though it was the smallest district. An average nearest neighbor spatial statistics analysis revealed that all incidents exhibited a random nature apart from Rape and Stealing which were statistically dispersed. None of the 8 crime types was found to be statistically clustered. The mean centers of the crimes were plotted and found to be within a 1 km radius.It is recommended that crime spatial data is collected over longer periods instead of a few months and should also include attribute data such as time, date and personal information of the perpetrators. The results of this analysis could be used as a predictive and tactical tool by the Ghana Police Service.

Keywords: Geographic Information Systems, Spatial data, Shapefile, Kernel Density Estimation, Spatial Statistics.