Spatial patterns analysis of urban road traffic accidents based on GIS

It is of vital importance to analyze road traffic accidents in order to improve traffic security management. Based on spatio-temporal analysis method, this paper aims to analyze traffic accidents data in time and space. Road traffic accident locations are usually described by geospatial coordinates, which can be mapped on road segments. Nevertheless, sometimes many traffic accidents locations were described by semantic data. Coordinates´ information has to be decoded from the semantic data before they are mapped on road segments. The paper firstly presents a solution to extract fuzzy locations from traffic accidents data. Secondly, this paper presents a spatio-temporal analysis method based on kernel density estimation (KDE) to extract the patterns of traffic accidents, namely, times, locations. Finally the traffic accident data of Beijing area are applied to test the feasibilities of the presented methods. The initial studies show that spatio-temporal analysis methods presented in this paper provide a powerful tool for the analysis of road traffic accidents.