Inspiration

There was a significant dog waste issue in the city this winter, enough for City Council to mount a waste reduction campaign. They passed the SCOOP Act on Thursday, August 13th. Discovering the "Canine Waste Dispensers" data set on the NYC Open Data portal during the Hackathon inspired the tool.

What it does

Enables users to report dog waste on the street and uses that crowdsourced data collection with 14 other NYC Open Data datasets to map hotspots and coverage gaps, enabling City to optimize deployment of resources.

How we built it

Used frontier AI coding model with most capability (Fable 5) to map out the plan and built product requirement document. Used lower models (Opus 5 and Sonnet 4.5) to build out the product features in an ~8 hour agentic loop. 45 git commits, 230 tests, 450M tokens, $225 equivalent API cost (a ~60% savings using tiered model routing).

Challenges we ran into

Deciding what to build!

Accomplishments that we're proud of

Getting a production-quality app deployed in about 12 hours! It's amazing what these tools can do.

What we learned

You'd better get everything built before you hit your weekly Claude Code limit...

What's next for Watch Your Step NYC

Let's get City Council to post about it in their dog waste reduction campaign

Built With

Share this project:

Updates

Submission history