"Big" Data as a Foundation for Measuring and Improving Public Transport Operations
Robert Bertini
Department of Civil and Environmental Engineering
Portland State University
Abstract
Over the past 20 years the transportation engineering field has witnessed a data revolution — some might say that we have transitioned from a data "desert" to a data "ocean." Intelligent transportation systems data can be archived and managed carefully to provide a platform for analysis, visualization and modeling. With increasing attention being paid to performance and financial issues related to the operation of public transportation systems, it is necessary to develop tools for improving the efficiency and effectiveness of service offerings. With the availability of high resolution archived stop-level bus performance data, it is shown that a bus trip time model and a bus stop spacing model can be generated and tested with the aim of minimizing the operating cost while maintaining a high degree of transit accessibility. In this research, two cost components are considered in the stop spacing model including passenger access cost and in-vehicle passenger stopping cost, and are combined and optimized to minimize total cost. A case study is conducted using one bus route in Portland, Oregon, using one year's stop-level archived Bus Dispatch System (BDS) data provided by TriMet, the regional transit provider for the Portland metropolitan area. Based on previous research considering inbound trips over the entire day, the theoretical optimized bus stop spacing was about 1,200 feet, as compared to the current value of 950 feet. Trade-offs will be discussed as well as an estimate of transit operating cost savings based on the optimized spacing. Given the availability of high resolution archived data, the paper illustrates that this modeling tool can be applied in a routine way across multiple routes as part of an ongoing service planning and performance measurement process.