This paper aims to show the usefulness of GIS and point pattern techniques for defining road-accident black zones within urban agglomerations. The location of road accidents is based on dynamic segmentation, address geocoding and intersection identification. One-dimensional (line) and two-dimensional (area) clustering techniques for road accidents are compared. Advantages and drawbacks are discussed in relation to network and traffic characteristics. Linear spatial clustering techniques appear to be better suited when traffic flows can be clearly identified along certain routes. For dense road networks with diffuse traffic patterns, two-dimensional techniques make it possible to identify accident-prone areas. The operationality of the techniques is illustrated by showing the impact of traffic-calming measures on the location and type of accidents in one Belgian town (Mechelen). (Author/publisher)
Samenvatting