WheelScore

Inspiration

Thinking about wheelchair accessibility- once a place is built not much can be done. What if there was a standardized metric for how wheelchair accessible a space is? A metric that could be computed even without a physical space, just a design?

What it does

WheelScore allows the user to place in a file (either a 3d file or an image) and takes the user to a page that displays a score and description/analysis of how wheelchair accessible the space is.

How we built it

We used the gemini API to handle images. For 3d files, initially we tried to use agentic systems to compute a score based on coverage, but for now we have hardcoded examples for object files.

Challenges we ran into

We initially tried to use agentic systems, but we felt we did not have adequate data (specifically enough 3d files) to simulate environments.

Accomplishments that we're proud of

We are proud we have a finished proof of concept site with sleek design.

What we learned

We learned that code implementation is a small portion of the process- ideating, brainstorming, implementing, it is an iterative cycle.

What's next for WheelScore

Our next steps are to use agentic systems to simulate device mobility in an environment and compute a score and analysis based on coverage.

Built With

• flask
• gemini-api
• next.js
• python
• react
• typescript