基于公交数据融合的城市人群移动模式识别——以重庆市中心城区为例

Movement Pattern Recognition of Urban Crowds Based on Public Transport Data Fusion:A Case Study of the Central Urban Area of Chongqing

揭示城市人群移动模式对理解人群移动行为、优化城市空间结构和解决交通问题具有重要意义。针对传统问卷方式存在数据量少、准确性差、实时性不足等问题,该文融合公交刷卡、公交GPS、公交线路和轨道交通刷卡数据,提出一种数据驱动的城市人群移动模式识别方法。首先,对多源公交数据进行匹配融合,提取公交、地铁联运的城市人群移动OD,再进行稳定居住地和稳定工作地识别,最后以稳定职住地匹配判别出行语义,识别居民通勤、购物、聚餐等社会活动出行模式。利用重庆市中心城区连续1周的公交数据进行验证,实验结果显示:①重庆市中心城区人群出行主要集中在工作日,且出行人数随出行频率的增加而降低;②79.34%的人群出行呈现“两点一线”移动模式,呈现三点、四点和多点移动模式的人群出行占比分别为13.03%、5.99%和1.64%。该文方法可以更全面精细地识别城市人群移动模式,有助于城市管理者、规划者了解城市人群移动方向,优化城市资源配置。

Revealing the movement pattern of urban crowds is crucial for understanding the movement behavior of urban crowds,optimizing urban spatial structure and solving traffic problems.In response to the problems of insufficient data volume,poor accuracy,and insufficient real-time performance in traditional questionnaire methods,in this paper,a data-driven method for identifying the movement patterns of urban crowds is proposed by integrating bus card data,bus GPS data,route basic data and rail transit card data.Firstly,the multi-source public transport data were matched and fused,and then the mobile OD of the urban crowds taking bus and subway combined transportation was extracted,and then the stable residential and workplace locations were identified,and finally the stable residential and workplace locations were used to interpret travel semantics,identifying mobility patterns for commuting,shopping,dining out,and other social activities.The method was validated using one week of public transportation data from the central urban area of Chongqing.The findings are as follows.①The crowd travel was mainly concentrated on weekdays,and the number of travelers decreased with the increase of travel frequency.②79.34%of the extracted urban crowd showed the movement pattern of"two points and one line",13.03%of that showed the movement pattern between three points,5.99%of that showed the movement pattern between four points,and 1.64%of that showed the movement pattern between multiple points.The paper proposes a method to identify more comprehensive and detailed urban crowd movement patterns,which is helpful for city managers and planners to understand urban population movement trends and optimize urban resource allocation.