基于轨迹数据的出租车运行模式识别及效益分析

Identification of Taxi Operation Patterns and Benefit Analysis Based on Trajectory Data

基于上海市出租车轨迹数据,使用Hadoop分布式计算框架对出租车GPS数据进行处理,以出租车为对象提取其轨迹,并计算出租车的载客平均距离、轨迹中心与城区中心距离与回转半径3个指标,采用k-means++模型分类并定义5种出租车运行模式。通过对比不同时间划分标准下各运行模式出租车的时间、里程载客率与收入情况,分析各类出租车在各时间段的运行效益与运行特征。

Based on the taxi trajectory data in Shanghai,this paper uses the Hadoop distributed computing framework to process the taxi GPS data,extracts the taxi trajectory as the object,and calculates the taxi’s average distance to carry passengers,the distance between the center of the trajectory and the center of the city and the radius of gyration,and uses the k-means++model to classify and define five taxi operation patterns.By comparing the operation time,spatial passenger load rate and revenue of each operation pattern taxis under different time division criteria,the operation efficiency and operation characteristics of each type of taxis in each time period are analyzed.