基于GIS分析的深圳市道路交通事故空间分布特征研究

掌握城市道路交通事故空间分布特征是城市道路交通安全管理的重要基础。基于深圳市2014~2016年的道路交通事故数据，首先应用地理编码方法对原始事故记录进行空间定位，形成事故的空间分布。其次针对考虑/不考虑路网密度的2种情况，应用密度分析方法对道路交通事故多发的区域和事故严重程度较高的区域进行鉴别，比较2种情况下区域分布的差异并分析造成这种差异的可能原因。最后利用异常点分析和热点分析2种空间聚类分析模型对事故严重程度较高的区域进行进一步鉴别，并对密度分析和聚类分析2种方法得到的结果进行了比较。密度分析结果表明：就事故频度而言，深圳市中心城区单位面积上的交通事故频度较高，而郊区单位长度道路上的交通事故分布更为密集；就事故严重程度而言，郊区的交通事故平均严重程度高于市中心区域。造成上述差异的原因可能与郊区道路限速较高等因素有关。聚类分析结果与密度分析结果相近，在郊区形成了高严重程度的事故聚类，而在中心城区形成了低严重程度的事故聚类，说明郊区的交通事故严重程度总体高于市中心区域。从2种方法的比较来看，密度分析简单易行，有助于交通管理部门对城市交通事故空间分布特征直观快速的了解；聚类分析可精确到事故点，为精细化的交通安全管理工作提供支撑。研究结果表明基于密度分析和聚类分析的研究方法对于确定道路交通事故空间分布特征有良好的作用。

GIS-based Spatial Patterns Analysis of Urban Road Traffic Crashes in Shenzhen

Understanding the spatial patterns of urban road traffic crashes is an important basis for urban road traffic safety management. Based on road traffic crash data from 2014 to 2016 in Shenzhen, China, this study first applied geocoding methods to locate the crashes and obtain their spatial distributions. Then, using the method of density analysis, areas with a high frequency and severity of crashes were identified. The differences in spatial patterns of crashes depending on whether road network density was considered in the density analysis were investigated and the possible causes of these differences were analyzed. Finally, using two types of disaggregated spatial clustering models, outlier analysis and hot spot analysis, those areas with a greater severity of traffic crashes were further identified. Comparative analysis was conducted on the results derived from the two clustering methods. The results show that in per unit land area, a greater number of traffic crashes occur in downtown Shenzhen than in the suburban areas of Shenzhen, whereas in per unit-length road, more crashes occur in the suburban areas of Shenzhen than in downtown Shenzhen. Severity of crashes in suburban areas is higher than that of the crashes in downtown areas. These regional differences may be related to the higher speed limits and inadequate traffic safety measures in suburban areas. The results of cluster analysis are essentially similar to those of the density analysis. In the suburban areas of Shenzhen, clusters of crashes with high severity are formed, whereas in the downtown areas, clusters of crashes with low severity are formed with low severity. This difference also means that the severity of traffic crashes in suburban areas is generally greater than that in downtown areas. A comparison of the two methods shows that the density analysis is simple and easy to perform and can thus help traffic control departments attain an intuitive and rapid understanding of the spatial patterns of urban road traffic crashes. Cluster analysis can be accurate to the level of crash points and can thus be used for delicate traffic safety management. The results of this study show that the methods of density analysis and cluster analysis can play a useful role in determining the spatial patterns of urban road traffic crashes.