Government Gains with the City of Boston

In our Government Gains series, we’re talking to dedicated public servants to learn three things about a recent project they’ve worked on that shows what’s possible when people ideate, collaborate, and innovate within government.

For this installment, we spoke with Anna Stoneman, who leads product development for the Citywide Analytics Team, and Michael Lawrence Evans, who directs the City of Boston’s Office of Emerging Technology. Their teams are collaborating on a project to map out the city’s curb space and figure out how best to distribute this finite space. Boston was recently awarded a SMART Grant from the US Department of Transportation to use machine learning to create a digital representation of the city’s street signs and curb regulations. With it, they’re building a tool to use large language models to extract regulation information from images of the street signs and store this in a standardized format. We spoke with them about what made this project challenging—and rewarding.

What was the biggest challenge you faced in this project?

Keeping track of curb regulations is really challenging. In some cases this information is just in people’s heads, in other cases our transportation planners have to check things with Google Street View—but no matter what, it’s a very manual process when we’re redesigning a street. If we want to allocate a space for five-minute delivery or plan for new modes of transit, we have people out there walking the street inch by inch and documenting every sign to get the data in alignment with our engineering drawings.

We applied for a grant with the US Department of Transportation to digitize all our curb regulations in order to help allocate curbs for parking, delivery, and outdoor dining—the full suite of things. One of the great things about this project has been the chance to learn from all the successes and challenges we’ve seen in other cities that have received similar types of curb management grants. We meet monthly with other cities in the SMART Curb Collaborative and they described the same issue: they hired a vendor to collect digital information on street signs, but they didn’t own the underlying technology to keep this information up to date and make it a tool that planners could use. 

Knowing that, we wanted to take a different approach: to do capacity building within our team rather than procuring this type of tech from a vendor. But that comes with different challenges.

We wanted to take a different approach: to do capacity building within our team rather than procuring this type of tech from a vendor. But that comes with different challenges.

How did you approach this challenge and how did you decide which tools to use to solve it?

When you want to own the tech for a project like this, you have to think long term. We have to scope out how we’re going to maintain it and the engineering resources we need to do that. There’s a trade-off in our approach compared to other cities because of that. To bring this work in house, we have to really carefully think about the skillset we need—and it’s challenging to hire in government, especially when that skillset is highly specialized! That’s meant we’ve had to pivot and are now figuring out how to work with a contractor who can build this technology with us.

It’s helpful that we’re leveraging two pieces of tech the city already has: an asset management system called Cartegraph, which stores all the signs that the city has installed. And we have an existing contract with Cyclomedia, a company that drives through our streets and provides panoramic imagery similar to Google Street View. We’re taking the data we have from those two products and trying to build a tool that integrates them both. 

Where will the lessons you learned here be applied in the future?

One thing we’re really trying to build on here is our use of open source data. We’re putting all our curb regulations in the same format as other cities—it’s a machine readable format that can easily be plugged into other technology stacks. And we’re hoping this makes it easier for other cities to adopt the platform we’re building. Having a simple, map-based interface that pulls in data sets with the curb regulations will be a welcome change. We hope we’ll get to a place where we can ask the system, “If we were to add a new loading zone in a business district, what would the effect be on parking availability, or traffic, or emissions?” Getting a better understanding of the cause and effect will make our planning efforts more effective—and we hope  other cities benefit from that type of technology, too.

For our own purposes, we’re hoping the work we do here really changes how we approach scenario planning and street modeling in the future. Once we have these base datasets, all that time we spend doing the manual work of documenting street regulations can be reallocated to interacting with people and figuring out the designs that work best. 

Want to learn more about Boston’s work on transportation technology? Check out their session at Code for America Summit—happening May 29-30, 2025 in Washington, DC.