🏙️ Mumbai House Price Predictor

A Streamlit web application that predicts residential property prices in Mumbai using a trained Decision Tree Regressor model (97.9% accuracy).

---

🚀 Live App

Deploy this app instantly on Streamlit Community Cloud:

1. Fork / push this repo to GitHub
2. Go to share.streamlit.io
3. Connect your repo → set Main file path to app.py
4. Click Deploy

---

📦 Project Structure

PricePrediction/
├── app.py                          # Streamlit frontend
├── requirements.txt                # Python dependencies
├── Mumbai-price-pridection.pickle  # Trained ML model
├── house_price_mumbai.csv          # Raw dataset
├── Price-Pridection.ipynb          # Model training notebook
└── .github/
    └── workflows/
        └── ci.yml                  # GitHub Actions CI pipeline

---

🛠️ Running Locally

# 1. Clone the repository
git clone https://github.com/manav-darji-aiml/PricePrediction.git
cd PricePrediction

# 2. Install dependencies
pip install -r requirements.txt

# 3. Launch the app
streamlit run app.py

The app will open at http://localhost:8501.

---

🧠 Model Details

Property	Value
Algorithm	Decision Tree Regressor
Accuracy (test set)	97.9%
Cross-val accuracy	~88% (ShuffleSplit, 8 folds)
Training samples	3,356
Features	Total sqft, Price/sqft, BHK, Location (one-hot)
Target	Price in Crore (₹)

Input Features

Feature	Type	Description
Location	Categorical	Mumbai suburb (20+ areas)
BHK	Integer 1–5	Number of bedrooms
Total Area	Integer (sq ft)	Carpet / built-up area
Price per sq ft	Integer (₹)	Local market rate

---

⚙️ CI/CD Pipeline

The .github/workflows/ci.yml workflow runs on every push/PR to main:

1. Install dependencies from requirements.txt
2. Verify all imports resolve correctly
3. Check the model file is present
4. Smoke-test the prediction logic end-to-end

---

📊 Data Preprocessing (Notebook)

1. Convert price strings (Cr / L) → float in Crore
2. Remove East/West suffixes from location names
3. Drop locations with fewer than 25 listings
4. Filter BHK types to standard 1–5 BHK Apartments
5. Remove outliers using BHK price-per-sqft statistics
6. One-hot encode Location (drop first to avoid multicollinearity)
7. Train/test split (80/20), GridSearchCV for hyperparameter tuning

---