“Smart” and “equitable” appear in the mission statement of nearly every public agency, consultancy, and start-up seeking to improve transit systems. However, as Ruha Benjamin writes in Race After Technology, “The road to inequality is paved with technical fixes.”
How, then, can transportation planners use data for equity?
It was my privilege to convene a discussion on this topic at the Hindsight Conference, an equity-focused urban planning summit hosted by the APA-NY’s Diversity Committee. The panel included Ambar Johnson of the LivableStreets Alliance, Jascha Franklin-Hodge of the Open Mobility Foundation, Lilly Shoup of Los Angeles’s Urban Movement Labs, and Sarah Williams, of MIT’s Civic Data Design Lab.
Our discussion underscored that transportation equity has never just been about mobility infrastructure. Data can, at best, capture only part of the story in complex cities. At worst, it can exacerbate the problem by giving planners and politicians a false sense of impact, without having addressed the root of transit inequity.
To advance equity through data, we must keep the focus on people.
Transit Equity is Complex
Graphics like this one suggest that smart mobility technologies — and, by extension, the data streams and algorithms that power them — will primarily benefit Black, indigenous, and people of color.
Some planners cautiously endorse this view, and suggested that “data is a tool for equity and inclusion.” Others go further, suggesting that comprehensive data can measure transit equity. As smart city tech gains traction in transportation and other sectors — observers increasingly ask, “Who wins when a city gets smart?”
The most optimistic view suggests that data can help overcome the social, political, and financial forces that have produced today’s inequitable transit systems. If so, planners must understand that transportation systems have always been political.
Consider the destruction of Black neighborhoods to construct the interstate highway system, or the Montgomery Bus Boycotts which were cued off by Rosa Parks’ refusal to cede her seat. Another landmark event — the voting rights march in Selma — traversed a bridge that was itself named for a white supremacist.
To use a more recent example, the Movement for Black Lives has built political power by occupying roadways. These protests have provoked shows of solidarity from some transit agencies, who refused to let sheriffs use public buses to transport protestors — as well as resistance from others, who offered their services to law enforcement, and even closed stations that demonstrators depend on.
To wit, transit systems are infused with social history. This history, in turn, shapes the data they produce and which inform algorithms used for transit planning. We must therefore understand data, and the way we use them, not just as a neutral descriptors of physical infrastructure — but as artifacts of political processes.
How, then, might planners use data to amend past and current injustices?
Engaging with historical inequities
A 2016 visualization by Massachusetts Bay Transportation Authority (MBTA) showed the bus routes with the most hours of passenger delay tend to serve Black and brown neighborhoods such as Roxbury and Mattapan. The same trend for a map of sidewalk quality conducted by the City of Boston: areas with crumbling infrastructure tend to be low-income and non-white.
Taken alone, these data can show which pieces of infrastructure are due for an upgrade. Viewing data through a historical lens reveals truths about power that purely quantitative analysis would not.
Ambar Johnson noted that many of these areas were subjected to the practice of ‘red-lining’ whereby federally-backed mortgage lenders would refuse fair financing to Black and brown residents who were deemed ‘high risk.’
“How many of us have seen the same kind of slides with the same blocks of where Black and brown folks live and where wealthy white communities are?” Johnson queried. “I think a lot of the times, we know exactly what’s going on with these historic implications.”
Indeed, histories of disinvestment have sowed doubt in marginalized communities as to whether cities will respond to their requests for assistance. This dynamic can lead to these communities being underrepresented in data, and cities to focus on communities with the greatest capacity to be seen — and less on those truly in need.
When the City of Boston compared the density of reported sidewalk complaints with an independent assessment of sidewalk quality in 2015, they found that complaints mostly came from white, wealthy communities — while areas with crumbling sidewalks were under-represented.
“It’s very much a cliché to say that what gets measured gets managed,” reflected Jascha Franklin-Hodge, who served as Boston’s Chief Information Officer at the time.
“By saying, ‘Hey, are we actually measuring the right thing?’ it gives us an ability to recognize that what we’re treating as objective data is heavily biased towards specific audiences that feel empowered to be able to demand services for their neighborhood.”
Aggregating data to the neighborhood scale, as in the graphic above, also introduces new pitfalls. Sidewalk quality doesn’t mean the same thing to everyone. Curb cuts may be especially important to people using wheelchairs, while sidewalk width may be more important to people with hearing impairments, who need to walk side by side to sign to one another.
Heat maps like this one obscure this nuance, and also fail to show how pedestrian networks complement public transit. Franklin-Hodge noted that a more effective approach would be to ask, “What are the key transportation paths and links that exist within specific neighborhoods? And in particular, the ones that facilitate access to public transit, since every transit journey starts and ends with a walking trip. “
Rethinking outdated models
In addition to recording injustices, data played a key role in creating them. Apart from racism, the key value underlying the red-lining maps was efficiency — which in this case facilitated the flow of wealth to federally-backed mortgage lenders and White homeowners, at the expense of Black and brown communities.
Similarly, well-intentioned efforts to “optimize” a city lead to unintended consequences today. Franklin-Hodge noted that, “inevitably, when you start to think about something as complex as city and the lives of the people who live in it, as something that can be optimized, you’re imparting your bias and your worldview about priorities into whatever system you build.”
When you start to think about something as complex as city and the lives of the people who live in it, as something that can be optimized, you’re imparting your bias and your worldview about priorities into whatever system you build.
Jascha Franklin-Hodge
Transportation planning’s roots in civil engineering can mean that influential standards, designed in an auto-centric era, often impede innovations for equity. For example, the “level of service” metric used in many transportation models emphasizes free-flow travel speeds, and ultimately encourages the design of high-speed roadways that only serve the need of motorists.
Validating these models, and the data that serve as inputs, against lived experience is critical. Lilly Shoup noted that a city she had worked with had implemented a travel demand model in the 1960s — which went uncalibrated for six decades, while it over-estimated auto travel demand by 5-10% year over year.
Ambar Johnson likewise suggested that “DOTs, or other city agencies do before and after counts of construction site of certain infrastructure projects, and compare those to whatever the model projected to what the actual reality is.“
Inaction through ‘smart-washing’
As Sarah Williams notes, data are people. However, data can often take people out of the picture, and enable planners to avoid painful — but critical — conversations about persistent injustice.
Discussing the comparison of the MBTA bus delay maps and the HOLC redlining maps, Johnson noted that the “the quality and the experience and the trauma that’s involved with those things that gets divorced with all the colors that get on the map.” In truth, cities have known where neighborhoods are in the most need of improved transit infrastructure for a last time.
“I’m afraid that we’re in a place now where people are more interested in acquiring the information and holding it and knowing what is wrong — but not doing anything about changing the material realities of the people who we are supposed to be serving.”
“I’m afraid that we’re in a place now where people are more interested in acquiring the information and holding it and knowing what is wrong — but not doing anything about changing the material realities of the people who we are supposed to be serving.”
Ambar Johnson
Franklin-Hodge calls this dynamic ‘smart-washing’ — a process through which cities avoid thorny issues that are fundamentally about power by using more data. As Chief Information Officer for the City of Boston, he observed “this tendency to look to technology in an attempt to make what are challenging political decision somehow easier or free of political friction.”
Painful histories are not likely to be resolved by the simple addition of more data. Rather, it takes concerted efforts to enact real change.
Does Data Lead to Action?
How, then, do we make sure that data leads to real, structural change in the built environment?
One way is to engage more people in the creation of data. Too often, Williams noted, “we don’t think about who or what is missing from the data,” which can lead to inadequate planning outcomes.
Working with semi-formal transit networks in Nairobi called Matatus, Williams’ team used cell-phone data to display the semi-formal network digitally — the same way one would view a subway or a bus network in Google Maps. Beyond making the system easier to navigate for people that rely on it, this process can itself engender new community.
Reflecting on the Digital Matatus project, Williams recalled how the drivers’ effort “to create evidence and advocate for their position actually helps bring them together and create that kind of community around a shared idea, or shared vision.”
Perhaps the answer is even more simple than that. Franklin-Hodge issued a challenge to transit officials: “If what you want is outcomes that really serve a community, maybe don’t start with data. Start with bringing the community together.“
Livable Streets embodied this approach by focusing on the busy often-delayed corridors that the MBTA identified. By simply talking with residents, they were able to quickly identify the intersections that posed the biggest challenges. As Johnson noted, “We were able to deliver those to the city and say, “Hey, this is a short term improvement list that you can get started on today.”
Indeed, Franklin-Hodge emphasized the “fine-grained knowledge and information that people are able to impart verbally. They can point to that one intersection and that one corner and say, you know, I keep almost getting hit here, because this thing happens, right? That is really useful data. It is not necessarily a quantifiable percentage of time somebody is almost hit, but you capture enough of those insights, and they actually can really serve the process in meaningful ways.”
Data can be layered in later on — but using that data should be driven by questions that are informed by ongoing inquiry, centered on lived experience. Data literacy programs such as that organized by the Boston Public Library can support citizens in accessing and making use of data to influence the planning process.
Cities are complex, which means there is no one-size-fits-all solution to create more equitable transit systems. Data can generate interest, maybe some new insights for equity, but people will always remain central to the planning process.
As Shoup put it, “we want to use data to set the table, rather than driving towards a solution.”
View the session recording here, along with the other sessions at Hindsight 2020.