Understanding transit travel behavior : value added by smart cards

Abstract

Travel behavior represents a particularly complex area of research in transportation given the interaction between transport supply characteristics and the user perceptions which guide his/her decisions. Thanks to the advent of automated data collection including smart cards, each customer transaction can now be recorded, providing a far more accurate, detailed and continuous stream of data for travel behavior analysis than the data previously available only from conventional travel surveys. This thesis explores the opportunities afforded by the data being generated by systems such as Automated Fare Collection (AFC), Automated Vehicle Location (AVL) and fare media in the form of smart cards, focusing on the Chicago Transit Authority and its smart card known as the Chicago Card. The processing of this data allows insights on transit user's decisions concerning the access to the transit system both at the origin and destination of the trip, together with the actual path choice and travel behavior dynamics. The knowledge gained on transit users' behavior could lead to important policy recommendations such as the value of consolidating or adding routes and services. In this context, the penetration of Chicago Card users among all system riders has been examined to determine how representative is the behavior of current card users. These card holders could also serve as a de facto longitudinal panel to gauge reactions to changes in the transit Level of Service (LOS). This thesis also prepares the ground for future path-choice modeling. The thesis presents an initial analysis of some basic travel parameters including frequency, time consistency, access distances, and route variability both for the first and the return trip among consistent users, as recorded during two weeks in September 2004 and September 2005.

(cont.) Given the exploratory nature of this thesis, several examples of actual trips have been visualized in order to develop some working hypotheses on travel behavior that might drive future research in this area. Finally, the thesis also discusses some of the potential data issues (reliability, synchronization and integrity between AFC and AVL records) involved in these analyses.