集时空聚类和指标筛选的公共交通通勤者识别  

Public Transportation Commuter Identification Based on Spatiotemporal Clustering and Index Screening

作　　者：周航 陈学武[2,3,4] ZHOU Hang;CHEN Xue-wu(Hangzhou City Planning&Design Academy,Hangzhou 310020,China;Jiangsu Key Laboratory of Urban ITS,Nanjing 211189,China;Jiangsu Province Collaborative Innovation Center of Modern Urban Traffic Technologies,Nanjing 211189,China;School of Transportation,Southeast University,Nanjing 211189,China)

机构地区：[1]杭州市规划设计研究院,杭州310020 [2]东南大学江苏省城市智能交通重点实验室,南京211189 [3]东南大学现代城市交通技术江苏高校协同创新中心,南京211189 [4]东南大学交通学院,南京211189

出　　处：《交通运输工程与信息学报》2022年第1期89-97,共9页Journal of Transportation Engineering and Information

基　　金：国家自然科学基金重点项目(51338003)。

摘　　要：通勤者作为公共交通乘客构成的核心部分,其识别提取是此类人群特征分析的前提。本文基于南京市常规公交、轨道交通和公共自行车的刷卡与设施数据,进行公共交通通勤者识别。首先,根据数据信息是否完整,分别采用两步聚类法和线路相似性整合法提取相似性出行;然后,识别职住地,再通过出行天数、单次出发时间差和工作往返出发时间差3项指标完成筛选。经通勤调查验证和方法有效性比较,各类参数取值合理,方法有效并存在应用优势。本文提出的通勤识别方法将出行时空规律与指标筛选紧密结合,考虑了数据完备与不完备条件下的不同数据处理思路,方法通用性和操作性强,识别结果能够为公共交通通勤乘客特征分析提供数据基础,有效指导后续城市公共交通设施布局和服务优化。Commuters are the core group of public transportation passengers,and their identification is the premise of these passengers’feature analysis.Based on the smart card and facility data of Nanjing’s bus,metro,and public bicycles,this paper proposed a method of public transportation commuter identification.First,two methods of two-step clustering and route similarity integration were applied to extract similar trips according to whether the data information was complete.The screening was then completed by home-work locations identification and three indicators:number of travel days,single departure time deviation,and work round-trip departure time deviation.After commuting survey verification and method validity comparison,the parameter values are reasonable,and the method proved effective and presents advantages in application.The identification method proposed in this paper combines the spatio-temporal pattern of regular trips and index screening and takes into account different data processing ideas under the conditions of complete and incomplete data.Therefore,the method has strong universality and operability.The results can provide a data basis for the analysis of public transport commuter characteristics and effectively guide the layout and optimization of future public transport facilities and service optimization.

关 键 词：公共交通 通勤识别 时空聚类算法 通勤者 多源数据 相似性出行 

分 类 号：U491[交通运输工程—交通运输规划与管理]