Competing or complimentary: Modeling transit ridership at route-level considering inter-route interdependencies

Transit routes in a network are often interdependent. Two or more transit routes complement or compete based on their spatial and temporal overlaps. Two routes complement each other if riders get to transfer from one route to another at a station. Two routes compete if they serve a common geographic stretch, giving the riders an option of choosing either one of them for their travel. When a new route is introduced by the transit agency, ridership on the existing routes changes too. Thus, it is important to assess inter-route relationships and their effect on transit ridership at route-level for better predictions. This research develops a route-level transit ridership model that is capable of capturing the effect of inter-route relationships in a transit network in addition to commonly used explanatory variables. Route-level ridership data were collected from the transit agency of the city of Gainesville, Florida (USA). For the city of Gainesville, model results showed that routes that share up to nine transit stops in a stretch have complementary relationship. However, when the number of common stops increases to 10 or more, routes start to compete. With every 10% increase in the supply of the routes that share ten or more transit stops with the subject route, logarithm of the monthly ridership on the subject route decreases by 0.96%.