Accident prediction models in urban areas

Abstract

According to official statistics, an important percentage of accidents and injuries are reported in Portuguese urban areas: from 2004 to 2007, 70% of the injury accidents and 43% of the fatalities occurred inside urban areas. To develop an efficient and affective strategy towards road safety, it is necessary that road administrations have the proper tools for the quantification of safety levels and the explicit consideration of safety issues in the road management process. The analysis of spatial accident distribution in road networks and the knowledge of the relations between accident frequencies and variables describing the urban road environment will allow a more efficient definition of priorities for intervention and safety funding. This can be achieved by means of accident prediction models adapted to the urban context where they are applied, and by the use of a Geographic Information System based methodology to analyse spatial patterns of road accidents. This paper summarizes the result of the bibliographic study on accident prediction models applied to urban areas and the data collection on road accidents, road infrastructure characteristics and traffic and land use information, integrated in a Geographic Information System, where it may be graphically visualized and analyzed.