Can taxi data inform bus route improvement? A case study in Shanghai

Taxis not only substitute for but also complement bus transit. With the complementary relationship, taxi trips reveal the origin-destination pairs where the level of bus service is low. This study attempts to utilize taxi data to identify potential weak links in the bus network. An evaluation methodology is proposed using taxi trips to evaluate the bus routes by incorporating accessibility measures at the stop level and convenience measures at the route level. With over 52,000 taxi orders collected in Jiading District of Shanghai, China, the corresponding alternative bus trips are simulated and then classified into 16 types of patterns according to the stop-level and route-level measures. The top five types, accounting for 78% of all trips, are selected for visual and quantitative analysis. The findings show that the proposed methodology can well assist bus agencies to improve service efficiency with better planning and design of bus routes.