Trash Mapper

Neighborhood cleanup work needs trustworthy location data. A simple open form can collect reports, but without validation it also collects irrelevant photos, duplicate noise, and reports that are hard to turn into action.

• An interactive map for browsing litter hotspots and submitted reports.
• A report flow with image upload, geolocation, severity, and cleaned-state updates.
• AI image validation that checks whether a submitted photo appears to show outdoor litter before accepting it.
• Supabase-backed authentication, Postgres persistence, and image storage.
• Image optimization with Sharp before storage, plus points for submitting and cleaning reports.

• Validate report images before they become trusted map data.
• Keep cleanup status separate from report creation so the map can show both problem areas and resolved work.
• Use user stats and points to make repeated civic contributions visible.
• Optimize uploaded images server-side so map browsing stays lightweight.

Public source and a deployed web surface make the project easy to inspect, run, and evaluate beyond a static screenshot.