Studies in the United States and Europe have demonstrated that burglary and vehicle crime exhibit consistent patterns, supporting the application of crime prediction techniques to proactively deploy police resources to reduce incidents of crime. Research into whether these techniques are applicable in an Australian context is currently limited.

Using crime data from the Queensland Police Service, this study assessed the presence of spatio-temporal patterns in burglary, theft of motor vehicle and theft from motor vehicle offences in three distinct local government areas. After establishing the presence of spatiotemporal clustering, the forecasting performance of two predictive algorithms and a retrospective crime mapping technique was evaluated.

Forecasting performance varied across study regions; however, the prediction algorithms performed as well as or better than the retrospective method, while using less data. The next step in evaluating predictive policing within Australia is to consider and design effective tactical responses to prevent crime based on the forecast locations and identified patterns.