Frequency Distribution And Influencing Factor Analysis Of Violation Behavior Based On Spatiotemporal Effect

According to the annual report on road traffic accidents,more than 240000 traffic accidents occurred in China in 2019,The domestic traffic safety situation is very serious.At present,the research on the frequency of traffic accidents mainly focuses on the micro and meso level,less from the macro level especially from the level of urban agglomeration.Taking the counties as the spatial analysis unit,this thesis uses the illegal data of Beijing Tianjin Hebei area to analyze the spatial and temporal distribution characteristics of speeding and drunk driving violation frequency;the significant influencing factors of speeding and drunk driving violation frequency are analyzed by constructing the traditional negative binomial model and spatiotemporal effect model of violation frequency.The results can provide a theoretical basis for local traffic management departments.The main work is as follows:(1)Spatial distribution and basic statistical analysis of violation frequency.The spatial distribution of speeding and drunk driving violation frequency is obtained by Arc GIS software,the results show that the frequency of drunk driving violations has a certain aggregation at the spatial level,mainly concentrating in the districts with large population density,more commercial activities and more entertainment places.Comparing the differences of violation categories of different vehicle types through cross contingency table,the results show that the main violation types of different vehicles are different,large vehicles have more overload violations and ordinary motorcycles have more violations without driving license.(2)Analysis of influencing factors of violation frequency based on negative binomial model.The negative binomial model of speeding and drunk driving violation frequency is constructed to reveal the influence of social and economic factors on violation frequency.The results show that the higher the area,population density,gross industrial output value and highway mileage,the more speeding violation frequency;the higher the per capita GDP,the less speeding violation frequency,the main reason is that the districts and counties with high per capita GDP are mostly urban core areas,which have large traffic flow and low speed and are not easy to exceed the speed limit.The higher the population,industrial output value above scale,highway mileage and road network density,the more drunk driving violation frequency,the main reason is that these districts and counties have more economic and commercial activities and more traffic flow,resulting in relatively frequent drunk driving.(3)Analysis of influencing factors of violation frequency based on spatiotemporal effect model.Four different spatiotemporal effect models of violation frequency are constructed by considering the temporal and spatial effect of violation in different districts and counties.The results show that the significant variable coefficient of the spatiotemporal effect model is larger compared with the negative binomial model,which indicates that the model without considering the spatiotemporal effect will underestimate the influence of independent variable on the violation frequency easily;the spatial error model(SEM)is the best model among the spatiotemporal effect models,which indicates that the spatial correlation of violation frequency mainly comes from the factors not observed in adjacent districts and counties.The results of direct,indirect and total effect of the spatiotemporal effect model shows that the violation frequency and independent variables of adjacent districts and counties have no significant effect on speeding violation frequency of their own districts and counties,while the independent variables of adjacent districts and counties have significant effect on drunk driving violation frequency of their own districts and counties,it mainly because the frequency of drunk driving violations is mostly concentrated in the core areas of the city,these areas are closely related to the commercial and economic activities of the surrounding areas,and the independent variables will interact with each other,but the traffic flow in these areas is more and the speeding phenomenon is less,which leads to the independent variables of similar areas have no significant impact on the speeding violations in this area.