Abstract: The need to study crime trends at smaller spatial units has long been recognized in the literature. Yet, researching patterns of crime at a micro-level of place i.e., individual buildings, addresses or street segments has been very slow in many developing countries such as Nigeria, perhaps due to the paucity of reliable data from the police incident reports. Data for this kind of research is required to be captured digitally and should contain the exact location of where incidents occur. In this paper, an approach is introduced for crime mapping at a micro-level of place, as well as for the estimation of crime clustering. This approach utilizes data that were drawn from a place-based victimization survey and a field mapping exercise. A total of 3,294 households, drawn from a sample of 12,524 residential homes, were interviewed of which 2,932 were included in the analysis. The utility of these data for geospatial and statistical analysis has been demonstrated here using a Nearest Neighbour Analysis method to show whether crime clusters at a micro-level of place, and if so, whether the clustering is more than would be expected on a chance basis. As demonstrated in prior studies, evidence is found to suggest that crime clusters in space more than would be expected in the study site, and the pattern is beyond mere chance. The implication of this finding for crime control and prevention is that strategies developed for elsewhere might also work in the settings of Nigeria. Future research could build on the findings from this study to further advance our understanding regarding the patterns of urban crime.
Key words: Crime analysis, Crime clustering, Crime mapping, Place-based victimization survey