Research On Accident Severity And Characteristics Of Urban Traffic Violations Based On Multi-Source Data

Road traffic injuries will not only threaten the life and property safety of traffic participants,but also cause huge economic losses to the society,which seriously affects the socio-economic development and social stability.The study of at-fault crashes severity and its influencing factors is an important research value for reducing traffic injuries and improving road traffic safety.The current rapid development of intelligent transportation and big data technology has brought many data support for the research of traffic safety and provided favorable conditions for exploring the at-fault crashes severity.Therefore,based on traffic accident data,traffic violation data and traffic safety awareness questionnaire data,this thesis scientifically evaluates the at-fault crashes severity and deeply explores relevant influencing factors from people,vehicles,roads and environment.The main research contents of this thesis are as follows:(1)Data screening,data cleaning,data transformation,data supplementation and other pre-processing operations on road traffic accident data and traffic violation data in Jinan City for three years from August 2016 to July 2019.Construct a multi-source traffic accident database,and use the number of fatal accidents,injury accidents and property damage accidents as accident level indicators,and the number of fatalities,injuries and direct property damage as accident severity indicators.Descriptive statistics on accident severity at multiple levels of people,vehicles,roads and environment,which lays the foundation for subsequent accident severity influence factor model construction.(2)Based on the characteristics of traffic accident data,this thesis proposes an accident severity grading method based on comprehensive accident intensity + K-means,which can comprehensively consider the accident consequences such as casualties,property damage and the type of traffic mode.Meanwhile,based on traffic accident data and traffic violation data,the accident severity evaluation index system is constructed.The final evaluation indexes are selected using correlation analysis,including the average accident severity,accident fatality rate and fatal accident ratio obtained from traffic accident data,as well as the accident severity rate and fatal accident ratio obtained from traffic violation data.On this basis,the entropy weight-grey association-TOPSIS method is selected to analyze the influence of various types of traffic violations on accident severity and judge the severity level of traffic violations,which can provide a theoretical basis for traffic management departments to carry out key violations regulation.(3)Medium and high-risk traffic violations are used as the research objects of accident severity influencing factors.Considering the factors of people,vehicles,roads and environment,13 factors are selected as independent variables from the traffic accident data.The results of comprehensive accident intensity + K-means classification and traditional classification are used as dependent variables to build a binary logistic regression model to identify the significant influencing factors of accident severity.Meanwhile,a multivariate logistic regression model is constructed based on the accident severity distribution characteristics of traffic accident data and traffic violation data.By supplementing the independent variables,16 factors are selected as independent variables from the traffic violation data and 13 factors are selected from the traffic accident data,and the results of accident severity by traditional trichotomies are used as dependent variables for impact factor analysis.Based on the results of traffic accident severity impact factor analysis,targeted improvement measures are proposed for significant impact factors,which can effectively reduce accident severity.(4)In order to explore the risk status of traffic safety awareness of travelers,a questionnaire survey is used to study the behavioral norms of travelers and their legal safety awareness characteristics.Reliability and validity analysis and factor analysis are performed on 1561 valid online questionnaires.The traffic safety awareness evaluation index system is constructed from three dimensions: traffic safety civilization cognition,traffic safety subjective norms,and traffic safety behavioral attitudes.The AHP-fuzzy comprehensive evaluation method is used to analyze human traffic safety risk factors from the perspective of demographic characteristics.Based on the results of public traffic safety awareness evaluation,effective traffic safety awareness improvement measures are proposed,which can reduce traffic accident risks from the human level.The research results of this thesis help traffic management departments to gain an in-depth understanding of road traffic safety risk factors and provide a reference basis for traffic violation management and accident prevention.