Using data mining techniques on APC data to develop effective bus scheduling plans

Various trip generators (e.g., buildings, shopping malls, recreational centers) continually influence travel demand in urban and suburban areas. As a result, the headway regularity that should be kept among transit vehicles is difficult to maintain, specifically during peak hours. The variation of headways lengthens the average wait times and deteriorates service quality. Providing a tool to monitor and maintain most up-to-date information through Advanced Traveler Information Systems (ATIS) can assist effective system planning and scheduling, while reducing the door-to-door travel time. This paper develops a methodology for clustering the state variables (number served passengers and halting stations in each vehicle trip) and using that for service planning. The data used to develop the models were collected by Automatic Passenger Counters (APC) on buses operated by a transit agency in the northeast region of the United States. The results illustrate that the developed tool can provide suggestions for improving systems performance as well as future planning.