Background

Access to safe, functional sidewalks is critical for building sustainable, equitable, and inclusive communities. Yet, in many parts of Washington, data about the pedestrian environment—such as sidewalk availability, condition, and connectivity—is incomplete or unavailable. This data gap hinders city planners from setting concrete project goals or evaluating environmental and accessibility impacts. It also discourages the public from choosing active transportation options.

Under the directive of the Washington State Legislature and in partnership with Gaussian Solutions LLC, the Taskar Center for Accessible Technology at the University of Washington is creating a comprehensive, high-quality pedestrian network dataset that includes sidewalks data across the state.

The project supports equitable mobility by equipping state agencies, local governments, and community organizations with the data needed to plan Complete Streets and prioritize resources effectively. Additionally, it aligns with Vision Zero safety initiatives by enabling data-driven decisions to prevent pedestrian injuries. Updating and expanding this dataset will empower Washington to provide all residents—especially vulnerable populations—with safe, accessible, and sustainable transportation options.

The Challenge

Mapping pedestrian infrastructure (sidewalks, crossings, curbs, traffic islands, etc.) for a state as large as Washington is a monumental task. Spanning approximately 185,000 square kilometers, the state’s pedestrian network is vast. Even discounting uninhabited forest areas, the target mapping area is still 8,550 square kilometers. Mapping such a large area manually would require an immense amount of time and manpower—an approach that is neither feasible nor cost-effective.

The Solution

Gaussian Solutions in partnership with the Taskar Center for Accessible Technology at Paul G Allen School of Engineering in University of Washington, Seattle have developed an AI-powered pedestrian infrastructure mapping solution. This solution accelerates the mapping process by more than 100x compared to manual efforts.

Using an AI-assisted, human-in-the-loop approach, the team generates high-quality, detailed, and connected pedestrian network data. This data is invaluable to transit agencies and city planners, enabling a wide range of analyses and applications.

The Technology

We developed Pedestrian Routing Infrastructure and Street Mapping AI (PRISM-AI) to solve the challenging task of mapping the missing key link in a complete street and transportation analysis: pedestrian network including sidewalks, crossings and curbs. PRISM-AI is a deep neural network based AI model that uses advances in computer vision and AI to ingest aerial imagery and produce high quality routable graph network of pedestrian infrastructure. Read more about the technology here.

The Data

We have completed the baseline mapping of the entire State of Washington that resulted in 39000km of sidewalks, 6700km of crosswalks, and 747000 curb locations. Sidewalks are enhanced with incline, surface and width information. You can view the data here

All the datasets are released via Transportation Data Equity Initiative(TDEI). Read about the TDEI project more here. All datasets are accessible by signing up in TDEI portal.

What’s next

We are putting final touches to the datasets ensuring all the data and the metadata is correct and is easily consumable by downstream applications. We are looking into expanding the elements we capture in the pedestrian infrastructure and to include the bike infrastructure as well.

Stay updated on our progress by following our LinkedIn page: Gaussian Solutions LLC on LinkedIn.

If you are interested in accessing this data, reach out to us at info@gaussiansolutions.com. We’d love to hear about your projects and how our detailed sidewalk data can support them.