Digital accessibility has been a key focus for governments as they work to meet a 2026 deadline to comply with a federal mandate. Some experts have argued that better data can better support people with disabilities by informing investments in things like accessible sidewalks and curb ramps. These investments would also support the curb cut effect, which is the idea that making things — like sidewalk curbs — more accessible for people with disabilities benefits the broader population. Implementing curb ramps has also made sidewalks more usable to older adults, people with baby strollers, or people with luggage.
In Lawrence, Kan., the state’s sixth largest city by population, officials are improving accessibility using GIS technology to support curb accessibility and compliance with the Americans with Disabilities Act (ADA).
“This really impacts a lot of people,” said Jessica Mortinger, transportation planning manager for the Lawrence-Douglas County Metropolitan Planning Organization within the city, calling it an equity issue as everyone is a pedestrian at some point. While at least 12 percent of people have a mobility issue, she noted that rate increases with age.
Lawrence, Kan.
A technical solution supports a strategic, systematic improvement process, Mortinger said. Using the Esri ArcGIS Network Analyst tool allowed staffers to compile data layers illustrating demand and ramp condition data. The map allows the city to visualize priorities, offering public transparency and enabling more efficient investment decisions.
In addition to improving the resident experience of the city’s pedestrian routes, using technology for ADA compliance planning protects cities from lawsuits, Korynta said.
The storytelling made possible through data illuminates the importance of this work, Montinger said, as people are “fighting for justice in being able to access their community.” It also would seem that it justifies the investment; while the city commission recently made budgetary cuts, there were no cuts to the $100 million ADA transition plan. It was passed in July 2024, so implementation work is just getting started.
“Part of the story is: we still have work to do,” Montinger said. Storytelling and the technology that supports it will be part of the road ahead, because “being able to quantify [the work] for the community over time is part of the responsibility as a steward of public dollars in this.”
Another locality taking on this work is Douglas County, Neb., which is working closely on GIS solutions with its largest city, Omaha. Both GIS and AI play a role in curb management.
About three years ago, Steve Cacioppo, the county GIS Department’s senior GIS analyst, started exploring deep learning tools and geospatial AI capabilities within ArcGIS to determine whether officials could create a model by extracting features from aerial imagery to inventory its more than 30,000 curb ramps.
Douglas County, Neb.
There were some initial challenges in proper identification when things like shadows, trees, or other objects obstructed them. The model initially considered some things, like a car sunroof or the painted rectangles making up a crosswalk, as curbs, which required correcting. Experts argue that AI still requires human oversight. Cacioppo acknowledged that the model is not perfect, but said drawing all of the curbs by hand would be both monotonous and incredibly time-consuming. One employee manually doing this, according to officials’ estimates, would have taken more than 1,000 hours, he said.
Other municipalities in the county received data about where curbs are located, too, which could support their own asset management, ADA compliance and accessibility efforts.
The data supports field work for city of Omaha ADA inspection staff, Brett Kelly, an Omaha GIS technician II, said, indicating the map layers underpin compliance reporting. Previously, he said, officials relied largely on “paper and pen.”
Cacioppo predicts this AI technology use case could be applied to other county inventorying needs, like identifying swimming pool or sewer locations.
“Maybe the AI finds something that they missed,” Cacioppo said. “Any other asset that anybody wants [to inventory], we have the ability and some knowledge now to create a model.”
Because this is a “supervised AI,” he said, it was easier to build buy-in because the input and output were both controlled.
“I am not a programmer or a developer by any stretch of the imagination,” Cacioppo emphasized, noting that he was able to do this by leveraging the tools available to him — and that others could, too.
