| Abstract |
Shifting vehicle drivers to alternate modes is becoming a key focus of city planning groups. Key to understanding how to posit new transit opportunities requires a granular understanding of origin-destination travel demand. By using Mobiliti, a HPC simulation developed at Lawrence Berkeley National Laboratory that populates origins and destinations and simulates their use of the transportation network, that granular understanding can be achieved. This data can be used to understand how current and potential future transit routes serve regional demand and how those services can be improved, an increasingly important aspiration in the face of falling transit ridership, increasing congestion, and environmental concerns. To understand how existing transit routes serve the demand, an origin and destination falling within a quarter mile around bus and light rail routes were considered zero-transfer transit-feasible; the results of this rudimentary analysis were noteworthy, with simulated feasible ridership up to five times higher than actual ridership seen by the Santa Clara Valley Transportation Authority. Using OpenTripPlanner, we found the average transit trip time to be approximately three times the driving time for that set of transit-feasible trips. This points to a need to increase the operational efficiency of transit services that cannot compete with auto traffic. The second portion of the analysis focuses on understanding how the transit network can be changed to better accommodate the demand. By understanding the origin-destination pairs unserved by the current transit network, corridors of high demand can be visualized, identified. Public transit is key to cities, and this methodology uses novel data sources to inform the design and development of that oft-overlooked workhorse of urban mobility. © 2019 Association for Computing Machinery. |
