Estimating the Destination of Unlinked Trips in Transit Smart Card Fare Data

Document Type

Journal Article

Publication Date


Subject Area

place - north america, mode - bus, technology - passenger information, technology - ticketing systems, ridership - behaviour


Smart card, public transit, unlinked trips, destination estimation, kernel density estimation


Smart card automated fare collection systems have been effective for the collection of data about the travel behavior of users on public transit networks. Because some systems record only the boarding (origin) locations, a method is needed for estimating the alighting (destination) locations. Existing algorithms can estimate the destination for most trips. However, unlinked trips, which are not part of a trip chain during the day, are more difficult to analyze. The proposed improvement to the existing model for destination estimation, especially for unlinked trips, is based on kernel density estimation of the spatial and temporal probabilities of each destination. The Société de Transport de l’Outaouais, a medium-sized bus service near Ottawa, Ontario, Canada, provided data for a 1-month period in 2009 (908,303 total transactions). Existing algorithms can handle only 80.64% of the trips; the proposed method handles an additional 10.9%. These results are analyzed, and future research directions are discussed.


Permission to publish the abstract has been given by Transportation Research Board, Washington, copyright remains with them.