ForkFinder

A fullstack restaurant discovery app built with React, Node.js, and MongoDB. You give it your location, it pulls real restaurants from OpenStreetMap and ranks them based on what you actually like to eat.

---

What it does

• Detects your location and fetches nearby restaurants from the Overpass API (OpenStreetMap's query layer)
• Caches results in MongoDB so you're not hammering the OSM servers on every visit
• Shows restaurants on an interactive Leaflet map
• Ranks results using a weighted scoring system that accounts for cuisine match, rating, and price range
• Lets logged-in users save favourites and set preferences (cuisines, price range) that feed into the recommendation engine
• Filter and paginate through restaurants on the Explore page

---

Tech stack

Frontend — React 18, React Router v6, React Leaflet, Vite
Backend — Node.js, Express, Mongoose
Database — MongoDB Atlas (with a 2dsphere index for geo queries)
External API — Overpass API (OpenStreetMap)
Auth — JWT with bcrypt password hashing

---

Project structure

restaurant-app/
├── backend/
│   └── src/
│       ├── config/         # MongoDB connection
│       ├── controllers/    # Route handlers (auth, restaurants, user)
│       ├── middlewares/    # JWT auth middleware
│       ├── models/         # Mongoose schemas (User, Restaurant)
│       ├── routes/         # Express routers
│       ├── services/       # Overpass fetcher, recommendation engine
│       └── server.js
└── frontend/
    └── src/
        ├── components/     # Navbar, RestaurantCard, FilterBar, MapComponent, Loader
        ├── context/        # AuthContext
        ├── hooks/          # useGeolocation
        ├── pages/          # Home, Explore, RestaurantDetail, Login, Register, Profile
        ├── styles/         # CSS files
        └── utils/          # Axios instance, helpers

---

Getting started

You'll need Node.js, npm, and a MongoDB Atlas connection string.

1. Clone the repo

git clone https://github.com/your-username/forkfinder.git
cd forkfinder/restaurant-app

2. Set up the backend

cd backend
npm install

Create a .env file in the backend/ directory:

PORT=5000
MONGO_URI=your_mongodb_atlas_uri
JWT_SECRET=your_secret_key
JWT_EXPIRE=7d
NODE_ENV=development

Start the server:

npm run dev

3. Set up the frontend

cd ../frontend
npm install

Create a .env file in the frontend/ directory:

VITE_API_URL=http://localhost:5000/api/v1

Start the dev server:

npm run dev

The app will be running at http://localhost:5173.

---

How the recommendation engine works

The scoring formula is:

Score = (Cuisine Match × 0.5) + (Normalized Rating × 0.3) + (Price Compatibility × 0.2)

Cuisine match is binary — 1 if the restaurant's cuisine is in your preferences, 0 if not. Rating is scaled from the 1–5 range to 0–1. Price compatibility drops off gradually — an exact match scores 1.0, one tier off scores 0.6, two tiers scores 0.3.

If you haven't set any preferences yet, the engine defaults to neutral weights (0.5) across the board so the ranking still works.

---

API endpoints

Method	Endpoint	Auth	Description
POST	/api/v1/auth/register	—	Register a new user
POST	/api/v1/auth/login	—	Login and get JWT
GET	/api/v1/restaurants	—	List restaurants (filter + pagination)
GET	/api/v1/restaurants/nearby	—	Fetch restaurants near lat/lng
GET	/api/v1/restaurants/recommend	✓	Personalised ranking for logged-in user
GET	/api/v1/restaurants/:id	—	Single restaurant detail
POST	/api/v1/user/favorites/:id	✓	Toggle a favourite
GET	/api/v1/user/profile	✓	Get user profile + preferences

Rate limiting is applied across all /api routes — 100 requests per 15 minutes.

---

Notes

• The Overpass API has a 10-minute cache baked in. If you hit /nearby for the same area within that window, it returns from MongoDB instead of calling OSM again.
• Restaurant data from OSM is inconsistent in some regions — names and addresses may be missing for less-documented areas. The app falls back to sensible defaults where data isn't available.
• The .env files are not committed. Don't push your credentials.