WardWatch: turning photos and voice notes into a live map of what a city actually needs
Most civic complaints never become data. They sit in a chat thread, get forwarded twice, and vanish — while the same pothole is reported forty times and nobody can say which street is worst. WardWatch (formerly Communit
Most civic complaints never become data. They sit in a chat thread, get forwarded twice, and vanish — while the same pothole is reported forty times and nobody can say which street is worst.
WardWatch (formerly CommunityPulse) is an AI-powered civic intelligence platform connecting citizens, local government, and service providers. The loop: a citizen reports an issue with GPS coordinates, a photo, or a voice note. AI does the boring part. Everyone else sees a live map.
What the AI actually does
Reports flow through LangChain custom workers calling Gemini and OpenAI — four jobs in one pipeline:
- Categorization — what kind of issue is this?
- Priority scoring — how urgent is it relative to everything else?
- Duplicate detection — forty reports of one pothole collapse into one record
- Content moderation — keep reports readable
Deduplication is the part most reporting tools skip. It's what turns a complaint box into a dataset a city can act on.
From reports to a map
Processed issues land on real-time heatmaps with geospatial clustering (Leaflet, MapLibre GL, Deck.gl). City admins get a dashboard with analytics for their area, and access is role-based: citizen / admin / superAdmin / city_admin, so a resident and a city official never see the same controls.
The platform also includes communities, events, real-time chat, and web push notifications — it isn't only a complaint queue.
The unglamorous details
- Stack: React 19, Vite 7, Chakra UI up front; Node/Express 4, Mongo 8, Redis, Pinecone, Socket.io, and Bull queues behind it; Docker, Sentry, and Cloudinary for the plumbing.
- Payments: CamPay mobile money for city subscriptions, the ads marketplace, and promotions — cards are not the default in the market this is built for.
- Audience: Cameroon-first: citizens, local government, service providers.
Takeaway
A tool that only collects complaints creates a backlog. One that categorizes, scores, deduplicates, and maps them creates a work queue.
Try it live: https://wardwatch-taap.onrender.com
Feedback welcome — especially from anyone who has tried to get a pothole fixed.
Originally published by Dev.to AI. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.