Forecasting bus ridership using a “Blended Approach”

Abstract

Abstract As sources of “Big Data” continue to grow, transportation planners and researchers seek to utilize these new resources. Given the current dependency on traditional transportation data sources and conventional tools (e.g., spreadsheets and propriety models), how can these new resources be used? This research examines a “blended data” approach, using a web-based, open source platform to assist transit agencies to forecast bus ridership. The platform is capable of incorporating new Big Data sources and traditional data sources, using modern processing techniques and tools, particularly Application Programming Interfaces (APIs). This research demonstrates the use of APIs in a transit demand methodology that yields a robust model for bus ridership. The approach uses the Census Transportation Planning Products data, modified with American Community Survey data, to generate origin–destination tables for bus trips in a designated market area. Microsimulation models us a transit scheduling specification (General Transit Feed Specification) and an open source routing engine (OpenTripPlanner). Local farebox data validates the microsimulation models. Analyses of model output and farebox data for the Atlantic City transit market area, and a scenario analysis of service reduction in the Princeton/Trenton transit market area, illustrate the use a “blended approach” for bus ridership forecasting.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 7 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.