CRIME OF THE SEASON : A SPATIOTEMPORAL ANALYSIS OF SEASONAL INFLUENCE ON STREET ROBBERIES IN MALMÖ

Crime in hotspot policing is assumed to have stable spatial patterns neglecting temporal dynamics. This study sets out to investigate seasonal differences of street robbery patterns in Malmö, Sweden. The objective is to generate knowledge and explore possible advantages using a more time-sensitive predictive approach for interventions and preventions. Local Indicators of Spatial Association (LISA) and Risk Terrain Modelling (RTM) were applied to test potential temporal variation between the seasons in hotspot patterns and spatial risk factor influence as well as comparing the predictive ability of seasonal to year-long models. The analysis confirmed the existence of temporal variation and only minor overlapping in street robbery hotspot patterns across seasons. Seasonal predictive models differed in their selection of spatial risk factors. Contrastingly, the few risk factors that were shared over more than one season did not differ in their impact. While seasonal models were better at predicting hotspots in limited areas, the year-long model was better at efficiently identifying street robbery hotspots across the whole study area. Consequently, the inclusion of temporality is valuable and should not be overlooked when planning interventions or preventions in hotspot policing. A combination of multi-temporal predictive models may increase the efficiency and effectiveness of strategic, tactical, and operative measures in designated areas of varying size. Future research should build on these findings and provide more optimised predictive models that guide practical implementations for hotspot policing.