Community Bus Demand Characteristics Analysis Based On Smart Card Data And GPS Data

Priority development of public transit is an effective measure to relief traffic jam in large cities. When developing public transit, many cities focused on macro-level issues of public transit system and gave great impetus to the construction of mass rapid transit systems, while paid less attention on micro-level issues. Thus the "last mile problem" has become one of the major causes that limit service level of public transit. In recent years, some cities start to plan micro-level bus routes, which make community bus emerging. Obviously community bus is a new thing which differs from conventional bus, and existing correlational researches are just rare. Completely understanding structure features of community bus routes and passenger behavior characteristics of this new system possesses can guide planning works of community bus routes.Traditional survey methods for analyzing passenger behavior characteristics are with high time and labor cost. Besides, the accuracy of their survey results is doubtful because of limited sample size of subjects. Through fusing and mining both IC card data of passengers and GPS data of buses, this thesis studied and analyzed route features passenger demand characteristics of community bus. Compared with traditional methods, more subjects are covered, data are gained more easily, and time span of data is longer.Firstly, this thesis briefly analyzed community bus and its relationship with conventional bus. Secondly IC card system, GPS system and the data structure of both systems are introduced. Thirdly this thesis discoursed data fusion and mining procedures including pre-processing aboard-mapping alight-mapping and transfer-judgment. After the above steps, OD matrix of passengers’transit trips can be gained. Then we analyzed passenger demand in community bus and conventional bus to seek critical factors of operation. Finally taking community bus lines in Chengdu as case study, we explored influence factors of passenger flow and turnover rate through regression analysis. This thesis conducted cluster analysis according to time-distribution features of passenger flow, then evaluated those lines’reasonableness with key indicators. Based on analytical results above, this thesis proposed improvement suggestions for different kinds of lines. Conclusions of this thesis provided scientific basis for route planning and improvement.