🏡 Mumbai Real Estate Price Prediction Engine

A Machine Learning powered web application that estimates residential property prices in Mumbai. It uses Ridge Regression with a sophisticated data pipeline to analyze location, area, and bedroom count, providing users with instant, data-driven valuations.

🚀 Click Here to Try the Live App

---

📸 Screenshots

---

🧠 The Intelligence (How it Works)

Unlike basic calculators that use simple averages, this project implements a full Supervised Learning Pipeline:

1. Data Processing: * Cleaned a dataset of 5,000+ real Mumbai listings.

• Handled high-cardinality location data (300+ unique regions) using One-Hot Encoding.
• Scaled numerical features (Area, BHK) using StandardScaler to normalize distributions.

2. The Model:

• Algorithm: Ridge Regression (Linear Least Squares with L2 Regularization).
• Why Ridge? Chosen over standard Linear Regression to handle multicollinearity between location features and prevent overfitting.
• Performance: Achieved an R² Score of 0.84, significantly outperforming the baseline.

3. Deployment:

• Wrapped the model in a Streamlit web interface.
• Deployed via CI/CD pipeline on Streamlit Community Cloud.

---

🛠️ Tech Stack

• Language: Python
• Machine Learning: Scikit-Learn (Ridge, Pipeline, ColumnTransformer)
• Data Manipulation: Pandas, NumPy
• Web Framework: Streamlit
• Model Persistence: Joblib

---

📂 Project Structure

├── app.py                 # The main Streamlit web application
├── train_advanced.py      # The training script (Data cleaning -> Pipeline -> Model Save)
├── model_advanced.pkl     # The trained serialized model file
├── cleaned_data_v2.csv    # Processed dataset used for the Location Dropdown
├── requirements.txt       # Dependencies for deployment
└── README.md              # Documentation

---

⚡ How to Run Locally

If you want to run this project on your own machine:

1. Clone the repository

git clone https://github.com/[YOUR-USERNAME]/mumbai-house-prices.git
cd mumbai-house-prices

2. Install dependencies

pip install -r requirements.txt

3. Run the app

streamlit run app.py

---

🔮 Future Roadmap (PropShare)

This tool is the MVP for a larger vision: PropShare, a fractional real estate investment platform.

• Phase 1 (Done): Price Discovery Engine.
• Phase 2 (In Progress): "Undervalued Deal" Alert System using Anomaly Detection.
• Phase 3: Fractional Investment Marketplace.

---

👨‍💻 Author

Yogin

• LinkedIn: www.linkedin.com/in/yogin-langalia

---