Reconstitution of the journeys to crime and location of their origin in the context of a crime series. A solution for a real case study.

Dynamic studies in crime analysis usually use distance-decay models applied on isotropic surfaces. If the distance actually traveled within the context of a crime series can be estimated, another approach, independent of the widely discussed form of the distance decay, can be suggested. Besides, there is a need for studies on anisotropic space, especially in European cities, given their complex road network. Therefore, this study takes this aspect into account.

We use data of a real case provided by the Belgian Police Department. Five criminal events have been committed in a short period of time by individuals using the same stolen car before abandoning it. The milometer allowed the Police to estimate the mileage covered by the criminals. The purpose of the analysis is to map out all the possible journeys to crime in order to find the offender's hideout.

First, we generate cost surfaces propagated from the locations of the criminal events over the dense road network in a high-resolution raster file. This provides a distance value for all pixels of the potential paths to each crime. Then these distances are cumulated for all criminal events and the sum is confronted to the recorded mileage. This gives a restricted list of road sections from which all event locations can be reached under the constraint of the mileage covered. These results are then refined through a multi-criteria analysis using exogenous data, such as land covers. The small area finally identified contains indeed the hideout as confirmed by the Police Department.