✈️ Flight Cost Intelligence

Compare Indian domestic flights by ₹ per kilometre — not just total price.

🌐 Live Site	https://pawankushwahh.github.io/Flight_per_km_cost/
⚙️ Live API	https://flight-cost-intelligence-api.onrender.com
📁 Frontend repo	https://github.com/Rakshita-0206/Flight_Project/tree/main/Flight_Fronted
📁 Backend repo	https://github.com/Rakshita-0206/Flight_Project/tree/main/Flight_backend

---

📖 About

Most travellers compare total fares — but a ₹3,000 short-hop and a ₹3,000 long-haul are not equal value. Flight Cost Intelligence normalises every Indian domestic route to ₹/km so you can compare fairly.

₹ per km  =  Total Ticket Price (₹)  ÷  Route Distance (km)

Lower ₹/km = better value per kilometre flown.

The project is split into two repos:

Repo	What it does	Hosted on
Flight_per_km_cost (this)	8-page HTML/JS frontend	GitHub Pages
Flight_per_km_backend	Flask REST API + data files	Render

---

🏗️ Architecture

Browser (GitHub Pages)
        │
        ▼  fetch JSON over REST (CORS enabled)
Flask API on Render
        │
        ▼  loaded once at worker startup
CSV + JSON data files (~118 routes, 26 airports)

• No database — data lives in flat CSV/JSON files loaded into memory at startup
• Frontend auto-detects localhost vs. production and switches the API URL accordingly
• A background ping to /api/ping warms the Render server on every page load

---

🖥️ Pages

Page	File	What it does
Home	index.html	Quick compare, popular routes, live stats
Route Compare	compare.html	Multi-route ₹/km table, bar chart, Leaflet map
Price Predictor	predictor.html	Monthly price trends, best booking month
Route Finder	route-finder.html	All destinations from one origin, ranked by ₹/km
Route Optimizer	optimizer.html	Nearby airports, cabin classes, layover options
Cost Heatmap	heatmap.html	India map coloured by cost intensity
Visualizations	visualizations.html	Cheapest/priciest routes, city averages
FAQ	faq.html	Methodology, data sources, technical info
404	404.html	Friendly not-found page

---

📡 API Endpoints

All responses: { "success": true, "data": ... } or { "success": false, "error": "..." }

Method	Endpoint	Description	Used on
GET	/api/ping	Health check / server warm-up	All pages
GET	/api/airports	All airports with coords	All dropdowns
POST	/api/compare	Compare routes by ₹/km	Compare
POST	/api/predict	Monthly price trend for a route	Predictor
POST	/api/route-find	Best destinations from an origin	Route Finder
GET	/api/nearby-airports	Nearby airports by IATA code	Optimizer
GET	/api/class-layover	Cabin class & layover data	Optimizer
GET	/api/heatmap	Regional cost heatmap data	Heatmap
GET	/api/visualizations	Cheapest/priciest routes + city averages	Visualizations, Home
GET	/api/raw-compare-data	Enriched routes (CSV + JSON merged)	Home popular routes

POST /api/compare also returns skipped and not_found arrays for invalid or missing routes.

---

🚀 Run Locally

You need both repos running simultaneously.

1 — Start the backend

git clone https://github.com/Rakshita-0206/Flight_Project.git
cd Flight_Project/Flight_backend

python3 -m venv venv
source venv/bin/activate        # Windows: venv\Scripts\activate

pip install -r requirements.txt
python app.py

Backend runs at http://127.0.0.1:5000

2 — Serve the frontend

git clone https://github.com/Rakshita-0206/Flight_Project.git
cd Flight_Project/Flight_Fronted
python3 -m http.server 5500

Open http://localhost:5500 in your browser.

config.js auto-detects localhost and points the frontend at http://127.0.0.1:5000.

---

📁 Project Structure

Frontend (Flight_per_km_cost)

├── index.html
├── compare.html
├── predictor.html
├── route-finder.html
├── optimizer.html
├── heatmap.html
├── visualizations.html
├── faq.html
├── 404.html
└── assets/
    ├── css/
    │   └── main.css           # Design system (dark theme, responsive)
    ├── images/                # Hero + sub-hero photos
    └── js/
        ├── config.js          # API base URL + endpoint map
        ├── common.js          # apiCall, fetchAirports, formatters, share URLs
        └── images.js          # Route thumbnail paths

Backend (Flight_per_km_backend)

├── app.py                     # Flask app + all API routes
├── gunicorn_config.py         # 4 workers, port 10000
├── Procfile
├── render.yaml                # Render deploy config
├── requirements.txt
├── scripts/
│   └── generate_data.py       # Regenerates JSON from CSV
├── data/
│   ├── compare_data_new.csv   # Canonical route prices (UPDATE FIRST)
│   ├── merged_flight_data.csv # Airport coords + names (UPDATE FIRST)
│   ├── compare_data.json      # auto-generated
│   ├── trend_data.json        # auto-generated
│   ├── class_layover_data.json# auto-generated
│   ├── heatmap_data.json      # auto-generated
│   └── nearby_airports.json   # auto-generated
└── tests/
    └── test_api.py            # 12 smoke tests covering all endpoints

---

☁️ Deployment

Frontend → GitHub Pages

Push to main — GitHub Pages auto-deploys in 1–3 minutes.

git add .
git commit -m "Your message"
git push origin main

Live at: https://Rakshita-0206.github.io/Flight_Project/

Backend → Render

Push to the backend repo's main branch — Render auto-deploys.

git add .
git commit -m "Your message"
git push origin main

Render runs pip install -r requirements.txt then gunicorn -c gunicorn_config.py app:app.

Setting	Value
Health check	/api/ping
Port	10000
Python version	3.9.0

Coordinating deploys

You changed…	Push frontend?	Push backend?
HTML, CSS, JS, images only	✅ Yes	❌ No
config.js API URL	✅ Yes	❌ No
app.py or data files	❌ No	✅ Yes
Both UI and API	✅ Yes	✅ Yes (backend first)

---

🗄️ Data Management

Regenerate derived JSON (after CSV updates)

cd Flight_per_km_backend
python scripts/generate_data.py          # write all JSON files
python scripts/generate_data.py --check  # coverage report only

Replace dummy data with real scraped data

1. Update compare_data_new.csv and merged_flight_data.csv
2. Run python scripts/generate_data.py
3. Push to GitHub → Render auto-redeploys (~2–5 min)
4. Run pytest tests/ -v to verify

---

🧪 Tests

cd Flight_per_km_backend
pip install -r requirements.txt
pytest tests/ -v

12 smoke tests cover all API endpoints.

---

🛠️ Tech Stack

Layer	Technology
Frontend	HTML5, CSS3, Vanilla ES6+
Icons	Font Awesome 6.4
Charts	Chart.js 4.4.1
Maps	Leaflet 1.9.4
Backend	Python 3.9+, Flask 2.0, flask-cors
Production server	Gunicorn + gevent (4 workers)
Storage	No database — in-memory cache at startup
Frontend hosting	GitHub Pages
API hosting	Render (free tier)

---

🔧 Troubleshooting

Problem	Fix
Slow first load (30–60 s)	Render free-tier cold start — wait and refresh
"Could not load route data"	Check backend is running; verify config.js API URL
Dropdowns empty	/api/airports call failed — check CORS and API URL
Maps or charts broken	Open browser console; check CDN scripts loaded
CORS errors locally	Ensure flask-cors is installed in the backend

Tip: Point an uptime monitor (e.g. UptimeRobot, free) at /api/ping every 10 minutes to prevent Render cold starts.

---

📝 Notes

• All prices in INR (₹); cost_per_km in ₹/km
• IATA codes must be 3 uppercase letters — invalid codes are returned in skipped
• Predictor uses historical trend averages from trend_data.json — not a live ML model
• Country is hardcoded to "India" in /api/airports
• Home page falls back to /api/visualizations if /api/raw-compare-data lacks cost_per_km
• Distances use the Haversine formula (great-circle, accurate to ~0.3%)

---

👥 Team

Name	GitHub
Rakshita	@Rakshita-0206
Shalini	—

📍 Lucknow, India