Are shelters in place? Mapping the distribution of transit amenities via a bus-stop census of San Francisco

Abstract

Transit stops serve as crucial components of journeys for riders, but their condition is often left out of equity considerations. Two important empirical questions are what stop amenities, such as places to sit, clear signage, shelters for inclement weather, and unobstructed curbs are present, and how are they distributed across systems, which may reveal neighborhood or route-specific disparities. San Francisco, CA represents an ideal case for which to pursue this question, given it maintains a ‘transit first’ policy directive that mandates public space prioritize transit over private automobiles. An in-person census of 2964 street-level bus stops was conducted over three months, which finds that a majority of stops lack both seating and shelter of any kind, that route signage varies widely in format and legibility, and that roughly one third of all stops are obstructed by on-street parking, rendering them difficult to use and exposing riders to oncoming traffic. Stops in the city’s northern half are more likely to feature seating, shelter, and unobstructed curbs, whereas amenity “coldspots” nearly all lie within the city’s southern half. Stop amenities also vary sharply by bus route, such that routes with the longest headways (and thus waiting times) provide on average the least seating, shelter, and clear curbs. These three amenities – seating, shelter, and unobstructed curbs – are also present to a greater degree in Census tracts with higher shares of white residents. This census demonstrates that equity evaluations of transit must include stop amenities, which are often overlooked, can undermine transit’s attractiveness, and even compound long-standing imbalances in service quality for underserved communities. Furthermore, studies of this kind can inform where amenity upgrades should be prioritized, targeting those areas currently lacking in high-quality stops, and raising the minimum standard of stop amenities overall. Finally, given data collected in this census is almost entirely unavailable to riders within current trip-planning and wayfinding applications, this work raises the possibility of expanding transit-data standards to include amenity details.