Road traffic safety influences people’s lives and properties. Safety evaluation should not only consider whether an accident happens, but also the potential risks. To date, significant research efforts have been dedicated to exploring the factors affecting accidents and the assessment of risks. This paper develops a traffic risk index (TRI) for evaluating the potential risks during a specific period on urban roads. Features are extracted from driving data offered by navigation corporation, and categorized by road class and road type based on the significance analysis. The index is calculated by TOPSIS method. The results show the different TRI distribution under urban road classes and types. Time-varying risk map and some degree of correlation with accident frequency are also demonstrated. This study can help to assess and predict urban road traffic risks, as well as providing references for personal drivers and relevant decision-making of the traffic management department to prevent road traffic accidents.

CICTP