Abstract

To protect the rights and opportunities of people with disabilities, which includes 1.3 million adults with legal blindness in the United States, the Americans with Disabilities Act (ADA) bans discrimination on the basis of disability in employment, transportation, and public services. Visually impaired people (i.e., those who are blind or have low vision) often rely on public transportation such as buses and subways for their daily transportation needs. Following the passage of the ADA law, regional transit agencies have implemented regulations regarding the accessibility of transit infrastructures. Typically, the requirements include large-print signage at bus stops, braille and tactile information within transit stations, and stop announcements inside transit vehicles at main points, among others.
Since 2005, transit agencies in the United States have started to open their transit data following General Transit Feed Specification (GTFS) standard, so that both spatial and temporal transit data can be accessible from mapping programs such as Google Maps. By combining detailed street and pedestrian mapping data, it is now possible to plan routes, at given departure times, from point to point, rather than just between transit stops. This open data initiative greatly helps to improve the accessibility of public transportation for individuals who are visually impaired.
The spatial element of transit data mainly comprises the locations of bus stops. The accuracy of the open transit data, upon which routes are planned, is especially critical for people who are visually impaired. If there is a considerable gap between the actual location of the bus stop and the location to which an individual with visual impairment is guided by Google Maps or other navigation applications or apps that use the open data, there is a good chance that they may miss the bus. This issue has been reported by visually impaired bus riders who rely on smartphones for navigation.
To estimate the accuracy of bus stop location mapping, we surveyed 174 bus stops served by the Massachusetts Bay Transportation Authority (MBTA) across six towns in suburbs of greater Boston (Arlington, Bedford, Belmont, Burlington, Lexington, and Newton). A user's physical distance to the affixed bus stop sign was measured from the location where the “blue dot,” which identifies the real-time location of the individual holding a mobile device running the Google Maps app, aligned with the bus stop icon.
Of the 174 bus stop locations we surveyed, nine bus stops were not mapped on Google Maps, and physical signs were absent at four of the mapped locations. As for the other bus stops, the average location error was 17.5 yards (STD: 18; range: 0–104), with 23% of them being off by more than 26 yards, twice the length of a bus (see Figure 1).

Distribution of absolute mapping errors. Note. The horizontal axis shows the distance from the physical bus stop sign to the location where the stop was shown to be mapped in the Google Maps app. The dashed vertical line indicates a distance equivalent to two bus lengths. The mapped location was greater than two bus lengths away from the physical stop sign board in 23% of all surveyed locations.
It is important to note that there were only a handful of sheltered stops in our sample. Most of the bus stops we surveyed only had a sign attached to a post or a utility pole, without the benefit of a bus shelter, which provides a larger discernable waiting area for passengers. When waiting to board a bus at a stop that lacks shelter, it is important for a passenger to stand close to the signpost so they may be identified as a passenger by the bus driver. For example, we experienced a bus driver not stopping to pick us up when we were standing a mere nine yards away from the bus stop signpost.
Massachusetts Bay Transit Agency updates GTFS open data four times per year, and Google Maps appears to implement the newest GTFS information regularly, as indicated by a couple of bus stops being updated about two months after our survey. Based on our observations, the primary cause for large errors is the relocation of the actual bus stop signs for various reasons, including bus stop renovation, replacement of signposts, and street redesign. The GPS coordinates, however, were often not updated in GTFS when the signs were moved.
Our findings indicate that mapping errors may pose a serious challenge for passengers with visual impairments who rely on location service-based navigation apps. Although we only surveyed a small sample within the MBTA bus network, it is possible that similar mapping errors are widespread in other regions across the United States. These challenges to accessibility can have significant implications, including diminished independence, compromised safety, feelings of isolation, and social disconnection. Consequently, individuals with visual impairments may be compelled to opt for alternative transportation options at higher costs.
To address these issues, we propose that transit agencies, together with blind and low-vision stakeholders, take the following actions to ensure bus stops are more accessible to people with visual impairments:
conduct more surveys and accessibility audits nationwide regarding the mapping errors in GTFS; facilitate communication between disability advocacy groups and transit agencies to ensure inputs from organizations representing individuals with visual impairments urging better accuracy in open transit data are received by transit agencies; offer training programs and resources to transit agency staff members, drivers, and customer service personnel to enhance their understanding of the challenges faced by individuals with visual impairments and how to provide appropriate assistance; and deploy assistive technologies and enhance wayfinding designs to provide more inclusive and accessible transportation systems (e.g., the mobile computer vision app All_Aboard, which can help bus riders find physical bus stop signs using smartphone cameras).
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The All_Aboard application (app) cited in this commentary is released by the authors’ lab for free. There is no revenue from app sale or in-app advertisements.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
