‘Applying and Evaluating Risk Terrain Modeling to the Canton of Zurich, Switzerland’
Crime prediction methods have been developed to not only analyze crimes that have happened in the past, but also to gain insight into their potential development in the future. One of these approaches is Risk Terrain Modeling (RTM), which allows localizing places with a high probability for future crimes. RTM is based on criminogenic factors or risk factors that can make it more likely that a crime is committed. The master thesis of Milena Kocher, which was the result of a six months internship at the Cantonal Police in Zurich, applied and evaluated RTM for urban and rural regions in the Canton of Zurich, Switzerland. Together with her supervisor Dr. Michael Leitner, Milena enhanced RTM with a local regression method to account for spatial variations in the relationships between different crime types and their selected risk factors. Results of this novel RTM-enhanced approach performed slightly better than the original, traditional RTM method. Overall, 20-50 % of future crimes were correctly predicted for high to highest crime risk regions, covering 1-7 % of the entire study area. This information proofs valuable for tactical and strategical planning for the police.