JOSHODIN2019/lagos

Lagos Explorer — interactive civic/POI map for Lagos, Nigeria with AI-powered natural-language queries over a Neo4j knowledge graph # Lagos Explorer **Live:** lagos-explorer.vercel.app (frontend, Vercel) · lagos-explorer-api.onrender.com (backend, Render) — both on free tiers. The AI query interpreter runs on Groq's free-tier API in this deployment instead of local Ollama (see Stage 12 in `PROJECT_MEMORY.md` and `LLM_PROVIDER` below). An interactive civic/points-of-interest map platform for Lagos, Nigeria — browse hospitals, schools, transport, banks, and more on a live map; search; save places; report civic issues; ask questions in plain English and get answers pulled from a real knowledge graph; and an admin dashboard to moderate it all. Built across 11 stages — see `PROJECT_MEMORY.md` for the full roadmap and per-stage build/verification log, and `docs/` for architecture, API, and component references. ## Tech stack | Layer | Technology | Why | |---|---|---| | Frontend | Next.js 16 (App Router) + TypeScript + Tailwind CSS v4 | Map UI, search, panels, admin dashboard | | Map | Leaflet + react-leaflet + leaflet.markercluster | Rendering ~3,100 POIs without choking the browser | | Backend | FastAPI (Python) | REST API serving the graph to the frontend | | Database | Neo4j Community Edition | POI/Category/LGA graph; also holds users, saved places, reports | | AI | Ollama (`llama3.2:3b`), local & free | Stage 8's natural-language → Cypher query interpreter | | Auth | JWT (bcrypt-hashed passwords) | Signup/login, protected routes, admin allowlist | No paid services anywhere in the stack — everything runs locally. ## Prerequisites - Python 3.11+, Node 20+ - Neo4j — `brew install neo4j && brew services start neo4j` - Ollama — `brew install ollama && brew services start ollama && ollama pull llama3.2:3b` ## Setup ```bash # Backend cd backend pip3 install -r requirements.txt cp .env.example .env # then edit .env — see SECURITY.md before setting real values python3 ingest_data.py # one-time: loads data/*.csv into Neo4j # Frontend cd ../frontend npm install ``` ## Running ```bash # Termin …