Abstract
Stop-level transit elasticities with respect to service frequency are estimated and discussed. Public transportation ridership is typically studied at an aggregate level, with variables influencing ridership averaged over time and space for a metropolitan area. Understanding transit rider-ship at a finer temporal and spatial level is generally limited to mode choice models. Most aggregate analyses are unable to capture important effects at the parcel or block level. Such analyses also cannot account for variation in demand over time of day, an issue that has been addressed to some extent by time series modeling. Data for the Chicago, Illinois, transit system were used, and results suggest that aggregate analyses overestimate the effect of service frequency on demand. In the context of other disaggregate analyses, these findings suggest that walking quality results in distinct increases in ridership, even after accounting for land use and for population and other demographics. A headway elasticity of rider-ship was estimated to be −0.263 to −0.277, very similar to recent disaggregate analysis of New York City transit data. The case is made for a better spatio temporal understanding of transit ridership for the effective allocation of resources.