🏡 AI House Price Predictor

A hybrid AI application that predicts real estate prices using both numerical data (bedrooms, square footage, etc.) and visual data (house images). Built with Deep Learning (TensorFlow/Keras) and Streamlit.

🎥 Demo Video

Watch the Project Demo on YouTube/Drive

📸 Application Demo

Below is a screenshot of the application in action, accurately predicting a "HousePrice" .

🧠 How It Works

This project uses a Multi-Input Neural Network:

1. Visual Model (CNN): Processes images of the house to extract visual features (condition, style, etc.).
2. Numerical Model (Dense): Processes structured data like bedrooms, bathrooms, and square footage.
3. Fusion Layer: Combines both inputs to output a precise price prediction.

🛠️ Tech Stack

• Python
• TensorFlow / Keras (Deep Learning)
• Streamlit (Web UI)
• Pandas & NumPy (Data Processing)
• OpenCV (Image Processing)

📂 Project Structure

notebooks/          # Jupyter notebooks for training and experiments
sample_images/      # Images used for testing predictions
app.py              # Main Streamlit application
house_price_model_v3.h5  # The trained AI model
cleaned_dataset.csv      # Processed data used for training
requirements.txt    # List of dependencies

💻 How to Run Locally

1. Clone the repository git clone https://github.com/sumitsolanki8055/HousePriceAI.git cd HousePriceAI

2. Install dependencies pip install -r requirements.txt

3. Run the app streamlit run app.py

📊 Dataset

The model was trained on a real estate dataset containing housing features and corresponding images. The data was cleaned and preprocessed in the notebooks/ directory.

🔮 Future Improvements

• Deep Learning: Upgrade from Random Forest to BERT or LSTM for better text understanding.
• More Languages: Add support for predicting difficulty in C++ and Java specific contexts.
• User Feedback: Allow users to vote on difficulty to retrain the model dynamically.

---

Created by Sumit Solanki