Semantic Reasoning for Geolocalized Assessment of Crime Risk in Smart Cities

The increasing number of crimes affecting urban areas requires the adoption of countermeasures to tackle this problem from different perspectives, including the technological one. Currently, there are many research initiatives with the goal of applying machine or deep learning techniques leveraging historical data to predict the occurrence of crime incidents. Conversely, there is a lack of tools aiming at crime risk assessment, in particular, by supporting the police in conceiving what could be the crime incidents affecting a given city area. To this purpose, we propose the Crime Prevention System, a modular software application for qualitative crime risk assessment. This consists of an ontology of crime risk, a module to retrieve contextual data from OpenStreetMap, semantics reasoning functionalities, and a GIS interface. We discuss how this system can be used through a case study related to the Italian city of Syracuse.