Identifying Hazardous Road Locations: Hot Spots versus Hot Zones

Abstract

Traffic safety has become top priority for policy makers in most European countries. The first step is to identify hazardous locations. This can be carried out in many different ways, via (Bayesian) statistical models or by incorporating the spatial configuration by means of a local indicator of spatial association. In this paper, the structure of the underlying road network is taken into account by applying Moran's I to identify hot spots. One step further than the pure identification of hazardous locations is a deeper investigation of these hot spots in a hot zone analysis. This extended analysis is important both theoretically in enriching the way of conceptualizing and identifying hazardous locations and practically in providing useful information for addressing traffic safety problems. The results are presented on highways in a province in Belgium and in an urban environment. They indicate that incorporating the hot zone methodology in a hot spot analysis reveals a clearer picture of the underlying hazardous road locations and, consequently, this may have an important impact on policy makers.