Traffic Accident Time Series Analysis

Abstract

Traffic accident hidden behind a lot of factors, although the rapid development of China’s transportation infrastructure, traffic management level is increasing, but the development of social economy and daily living travel traffic caused by the supply and demand contradiction is prominent. The traffic safety management situation is still grim, needs to be based on existing management, through the analysis of massive traffic the management of data mining, finding the underlying changes in the law, for the transportation departments to effectively carry out the administration of road traffic safety, more in-depth mining the potential causes of traffic accidents. In this paper, we first use the Baidu map API interface to transform the street information to the specific latitude and longitude. Then we divide the morning and evening peak and non peak hours, working days and non working days, and analyze the time characteristics of the accidents based on the MATLAB 3D data imaging principle. Secondly, the stability analysis is carried out after the data collection, and the occurrence trend and periodic law of the accident are obtained by using the Holt - Winters filter prediction and test and STL decomposition. Finally, in view of the above analysis, the reasonable countermeasures are given in order to reduce the incidence of traffic accidents and build a harmonious travel environment.