🏡 Land Price Predictor

A production-ready Machine Learning + FastAPI project that predicts land price per cent for Kannur district (Kerala).

---

📘 Overview

The Land Price Predictor is a machine learning project designed to estimate the price of land (in lakhs per cent) based on real-world factors such as location, land area, distances to key facilities, and land type.

This project combines data science with a modern backend to provide real-time predictions.

🛠️ Tech Stack

• 🧠 Scikit-learn – Machine Learning (Random Forest Regressor)
• ⚡ FastAPI – High-performance backend API
• 🐼 Pandas – Data manipulation & preprocessing
• 🌍 Synthetic Dataset – Realistic data modeled for Kannur, Kerala
• 💾 Joblib – Efficient model persistence
• 🧰 JSON Config – Centralized parameter management

---

🗂️ Project Structure

server/
│
├── app.py                         # FastAPI main entry point
├── api.py                         # API route logic (/api/predict)
├── model_loader.py                # Utility to load model & preprocessing columns
├── schemas.py                     # Pydantic models for request validation
├── config.json                    # Project & Model configuration
│
├── src/
│   ├── preprocess.py              # Data cleaning and feature engineering
│   ├── train_model.py             # Script to train & evaluate the model
│   └── __init__.py
│
├── data/
│   └── land_price.csv             # Training Dataset
│
├── models/
│   └── random_forest_model.pkl    # Serialized ML model (generated after training)
│
└── requirements.txt               # Python dependencies

---

⚙️ Installation

Follow these steps to set up the project locally.

1️⃣ Clone the Repository

git clone https://github.com/Muhammed-Rizin/land-price-predictor.git
cd land-price-predictor

2️⃣ Install Dependencies

It is recommended to use a virtual environment.

pip install -r requirements.txt

3️⃣ Verify Dataset

Make sure land_price.csv exists in the data/ folder.

---

🧩 Configuration (config.json)

This file controls the project behavior. You can easily adjust model parameters without editing code.

{
  "data_path": "data/land_price.csv",
  "test_size": 0.2,
  "random_state": 42,
  "n_estimators": 100
}

---

🚀 Usage: Machine Learning Pipeline

🧠 1. Train the Model

Run the training script to process the data, train the Random Forest model, and save the .pkl file.

python src/train_model.py

✅ Output:

✅ Model saved successfully as 'models/random_forest_model.pkl'

📊 Model Evaluation:
Mean Absolute Error (MAE): 0.42
R² Score: 0.93

🔮 2. Run Prediction (CLI)

You can test predictions directly via the command line before running the server.

python src/predict.py

✅ Output:

===========================================
🏡 LAND PRICE PREDICTION RESULT
-------------------------------------------
Predicted Price per Cent: 2.73 lakhs
===========================================

---

🌐 Usage: API Server

1️⃣ Start the Server

Launch the FastAPI application using Uvicorn.

uvicorn app:app --reload

2️⃣ Access the Interface & Endpoints

Once the server is running, you can access the following endpoints:

Method	Endpoint	Description
GET	/docs	Swagger UI Interactive dashboard to test APIs directly in the browser.
GET	/redoc	ReDoc Alternative, clean documentation view.
POST	/api/predict	Prediction API Main endpoint to generate price estimates (accepts JSON).

---

📡 API Documentation

Predict Land Price

Endpoint: POST /api/predict

📥 Request Body (Example)

{
  "land_area_cents": 7.25,
  "land_type": "Residential",
  "taluk": "Kannur",
  "village": "Pallikunnu",
  "location_name": "Kannur City",
  "latitude": 11.8745,
  "longitude": 75.3704,
  "distance_to_school_km": 1.5,
  "distance_to_airport_km": 25.0,
  "distance_to_railway_station_km": 3.0,
  "distance_to_hospital_km": 2.0,
  "distance_to_medical_college_km": 5.0,
  "distance_to_bus_stop_km": 0.5,
  "distance_to_market_km": 1.0
}

📤 Response (Example)

{
  "price_per_cent": 2.73,
  "total_price": 19.79
}

---

📊 Dataset Summary

The model is trained on a realistic dataset containing the following features:

Field	Description
property_id	Unique ID for the property
location_name	Area in Kannur (e.g., Pallikunnu, Thalassery)
taluk	Sub-division under the district
village	Local village name
latitude, longitude	Approximate geolocation coordinates
land_area_cents	Plot area in cents
land_type	Residential, Commercial, Agricultural, or Mixed
distance_*	Distances to amenities (School, Hospital, Airport, etc.)
price_lakhs	Target Variable: Total land price (Lakhs INR)

---

🧮 Model Details

• Algorithm: Random Forest Regressor
• Training Split: 80% Train / 20% Test
• Performance:
  • High R² Score indicates strong correlation.
  • Low MAE ensures price estimates are close to actual market values.

---

🧑‍💻 Author

Muhammed Rizin 📍 Kerala, India 💻 Full-Stack Developer 🔗 GitHub → Muhammed-Rizin 🔗 LeetCode → muhammed-rizin

---