基于位置数据的居民出行时空特征研究——以上海市为例

Research on temporal and spatial characteristics of residents’ travel based on location data——A case of Shanghai

为了有效提取城市居民出行特征,结合出租车轨迹数据和POI数据,以上海市为研究区域,采用一种基于DBSCAN和K-means的混合聚类模型对上海市POI数据进行空间聚类分析,计算出区域集聚和中心后,引入泰森多边形(Voronoi图)的概念,将城市划分为多个人群聚集区域;通过对轨迹数据的分析,挖掘居民出行的时空特征,并对其成因做出了简要的分析。实验结果表明:居民出行特征具有明显的随时间变化的规律,各区域的出行高峰时间基本一致;居民出行特征在空间上呈现出明显的距离衰减效应,区域间距离越远,交互强度越弱。

In order to extract the travel characteristics of urban residents effectively,Based on the taxi trajectory data and POI data,this paper uses Shanghai as the research area,and uses the hybrid clustering model based on DBSCAN and K-means to spatial cluster analysis of Shanghai POI data.After calculating the regional agglomeration and center,the concept of the Voronoi diagram is introduced to divide the city into multiple crowded areas,Through the study of trajectory data,the spatial and temporal characteristics of residents′travel are excavated,and the causes are briefly analyzed.The results show that the residents′travel characteristics have obvious time-varying regularity,and the travel peak time of each region is basically the same.The residents′travel characteristics show obvious distance attenuation effect in space.The farther the distance between regions is,the weaker the interaction intensity is.