BusTr

Abstract

We present BusTr, a machine-learned model for translating road tra c forecasts into predictions of bus delays, used by Google Maps to serve the majority of the world's public transit systems where no o cial real-time bus tracking is provided. We demonstrate that our neural sequence model improves over DeepTTE, the state-ofthe-art baseline, both in performance (−30% MAPE) and training stability. We also demonstrate signi cant generalization gains over simpler models, evaluated on longitudinal data to cope with a constantly evolving world.

Review of Transit Data Sources: Potentials, Challenges and Complementarity

“…[125] study the problem of estimating travel times on public transport buses with real-time traffic information. Google researchers are also interested in this topic, as their recent work ( [126]) shows.…”

“…Another option is to use the open data tools available, such as GTFS, which could be a very useful way to manage the storage budget. Although agency involvement is required to provide these data and invest in the underlying technology, research continues to facilitate the use of these tools (e.g., [126]).…”

“…The problem with AVL data is that it is usually not yet openly available for the majority of transit data. In [126], the authors integrate (predicted) traffic data as a replacement of GTFS where they are unavailable. In [195], the authors investigate the effects of vehicle delays on passenger waiting time together with the effects of transfer status, boarding location, time of day and rider travel frequency.…”