The value of automated fare collection data for transit planning : an example of rail transit OD matrix estimation

Abstract

Traditionally, transit agencies across the world have relied on traveler surveys and manual counts to inform many of their service and operations planning decisions. Today, many agencies can add to their existing planning toolbox the data obtained from new Automated Fare Collection (AFC) technologies. By adding this dataset, transit agencies can boost their analytical capabilities and deal with some planning questions that they previously could not easily address. In fact, while with surveys and manual counts transit agencies were able to form a reasonable snapshot of existing demand on their transit system, with accurate AFC data, planners should be able to get a detailed, continuous and accurate vision of the travel behavior of their customers, at a fraction of the prior cost. Nevertheless, there are some technical and operational issues that can affect the quality of AFC data that must be addressed before the new dataset can be fully integrated into the planning process of transit agencies. This research begins to explore these issues in general as well as in the context of the transit system serving London in the United Kingdom. In particular, it identifies bias in the AFC entry and exit data and develops a methodology for building an unbiased estimate of existing travel patterns on the London Underground.

(cont.) The outcome of the research is a methodology to build unbiased estimates of existing travel patterns. The use of this methodology presents two main advantages over the existing survey methods: (i) the resulting estimate corrects the bias in the Oyster dataset and better reflects existing travel patterns than the traditional survey-based methodology and (ii) the methodology should be easy to replicate, offering planners the capability to build origin - destination matrices specific to different time periods, days of week and seasons of the year. The availability of this large set of origin - destination matrices should enable planners to keep track of changes in travel patterns and tackle many planning questions that they could not easily address before.