Ridership effects of real-time bus information system: A case study in the City of Chicago

Abstract

In this paper, using longitudinal data on route level monthly average weekday ridership in
\nthe entire Chicago Transit Authority (CTA) bus system from January 2002 through December
\n2010, we evaluate the ridership effects of the CTA real-time bus information system.
\nThis bus information system is called CTA Bus Tracker and was incrementally implemented
\non different CTA bus routes from August 2006 to May 2009. To take account of other factors
\nthat might affect bus ridership, we also include data on unemployment levels, gas prices,
\nlocal weather conditions, transit service attributes, and socioeconomic characteristics during
\nthe study period. This combined longitudinal data source enables us to implement a
\nquasi-experimental design with statistical controls to examine changes in monthly average
\nweekday ridership, before and after the Bus Tracker system was implemented, on each bus
\nroute. Based on a linear mixed model, we found that the provision of Bus Tracker service
\ndoes increase CTA bus ridership, although the average increase is modest. Further, the
\nstudy findings suggest that there are temporal variations of the ridership effects among
\nthe routes, with the "winning" routes more likely to have the technology implemented
\nin the later phases of the overall "roll-out" period. However, the results are less conclusive
\nregarding geographical variations in the effects of Bus Tracker