User behavior transition mapping for bus transportation planning based on time series data analysis of travel e-ticket information: A case study in Tainan City, Taiwan

In the past, the bus service planning in the metropolitan area is an individual procedure of bus operators, private organization, or local government. Most of them use their own data and analyse them to build an appropriate planning to operate. One of the important topic of bus service improvement is that understanding users' decision. However, there are only rare users will announce their thought. In order to understand user's will, we use a data mining technique to obtain information from large e-ticket data. Last decade, the e-ticket system apply on many public transportation systems widely, and it is easy to get all transaction information of e-ticket users. In order to obtain user's decision based on e-ticket, we propose a method to cluster Behavior and conduct the Behavior transition calculation. In the Behavior transition map, manager can understand the overall transition tendency in whole system, and target clusters need to improve.