交通事故的时空热点分析

Spatial-temporal hotspots analysis on traffic accidents

针对目前交通事故分析中时空维度分离的不足,以H市2013-2015年的交通事故数据为研究对象,根据自组织神经网络、平行坐标系、时空颜色矩阵和时空网络核密度估计、热点分析法,分别从行政区划以及道路网络上进行交通事故的时空特征分析,从宏观和微观角度揭示交通事故的时空热点区域。结果表明,整体来说,在空间上,H市交通事故热点区域4个街道高于8个镇,呈明显的“两带一中心”分布,即硖许线与101省道形成的带状区域和市政府所在的行政中心区域;时间上,早、晚高峰时段最为严重,夜晚较为轻微,呈明显的区域特征分布,即工业办公区热点出现在早高峰时段,生活住宅区热点出现在晚高峰时段,商业消费区热点出现在夜晚时段。

Against the deficiency of spatial-temporal dimension separation in current traffic accident analysis,this paper takes the traffic accident data of H city from 2013 to 2015 jointly applies techniques of self-organizing map,parallel coordinate system,spatial-temporal color matrix and spatial-temporal network kernel density estimation,Getis-Ord Gi^*statistics,to analyse the spatial-temporal characteristics of traffic accidents based on the administrative divisions or on the road network,revealing the spatial-temporal hotspots of traffic accidents from the macro and micro perspectives.The results show that the hotspots area of traffic accidents in H city are distributed more in four streets than in eight towns,and have obvious spatial distribution of“two belts and one center”,the belt areas consist of Xiaxu line,101 provincial highway and the administrative center area where the municipal government is located.For temporal distribution,the hotspots are most serious in the traffic rush hour and slightly at night,and obviously correlated with regions,that is,hotspots in industrial office areas occur in morning peak period,hotspots in residential areas occur in evening peak period,hotspots in commercial consumption areas occur in night period.