I. INTRODUCTION
With the rapid development of our national economy, traffic safety issue becomes increasingly prominent, which has attracted widespread attention. Data mining in traffic accidents, which helps to find the hidden knowledge and rules has become an important research area in traffic safety. Currently, most of the traffic information analysis is limited to general statistical analysis, which is hard to explore the rules hiding in traffic accident information. Also, they do not have the capability of map displaying and spatial analysis, so as not be able to find the spatial distribution characteristic and relationship between traffic accidents and road network elements[1]. In recent years, GIS has been developed rapidly and used broadly in the field of traffic safety. In the developed countries, especially US and Western Europe, GIS technology has been applied to urban traffic information management widely. However, in our country, researches of GIS technology application in traffic safety area were started rather late. The public transport information system developed by the Ministry of Public Security is representative in domestic traffic safety systems and it is even not combined with GIS totally. The research groups in Jilin University and Tongji University have made discussion about project level traffic safety evaluation information system based on GIS. In the meantime, some domestic universities and research institutions have also made a lot of researches of applying GIS technology to traffic safety area[2] [3]. Anyway, domestic research and application of traffic accident information system based on GIS is still in the infancy and a lot of problems need to be solved. In the paper, it is firstly discussed how to utilize GIS technology to locate the traffic accidents which are described as address with text in road network. Secondly, based on the spatial relationship between traffic accidents and road network elements, black spots are extracted. At last, the reasons for the accident-prone spots are detected, which helps to provide decision supports for traffic safety.