AI-Powered Real Estate Price Predictor

📖 Overview

This project is a complete, end-to-end MLOps system designed to predict housing prices in Ames, Iowa. It follows a modern software engineering and machine learning workflow, encompassing everything from data exploration and model training to API development, containerization with Docker, and automated deployment to the AWS cloud.

The core of the project is a Gradient Boosting (XGBoost) model, a high-performance algorithm renowned for its accuracy on tabular data.

---

📊 Model Performance

The model was evaluated on an unseen test set (20% of the original data). The performance metrics indicate a highly accurate and reliable model.

• R-squared (R²): 0.9037
  • This means the model can explain approximately 90.4% of the variance in the house sale prices.
• Root Mean Squared Error (RMSE): 0.1341
  • This is the average prediction error on the log-transformed sale prices.

Metric	Score	Description
R-squared (R²)	0.9037	Explains 90.4% of price variance.
RMSE (on log scale)	0.1341	Low average prediction error.

---

🛠️ Tech Stack

Component	Technology/Library
Data Science	Pandas, NumPy, Scikit-learn
ML Model	XGBoost
Backend API	FastAPI
Web UI	Streamlit
Containerization	Docker
Testing	Pytest
Automation	GitHub Actions
Cloud Deployment	AWS (Elastic Beanstalk, S3, ECR)
Version Control	Git & GitHub

---

🚀 CI/CD Automation

This project is configured with a complete CI/CD pipeline using GitHub Actions. On every push to the main branch, the following automated workflow is triggered:

1. Run Tests: The pytest suite is executed to ensure code quality and prevent regressions.
2. Build & Push: A new Docker image is built and pushed to a private Amazon ECR repository.
3. Deploy: The new image is automatically deployed to the AWS Elastic Beanstalk environment, updating the live application.

---

🚀 Getting Started

To run this project locally, follow these steps:

Prerequisites

• Python 3.11+
• Docker Desktop

1. Clone the Repository

git clone [https://github.com/Murci20965/real_estate_price_predictor.git](https://github.com/Murci20965/real_estate_price_predictor.git)
cd real_estate_price_predictor

2. Set Up the Environment

# Create and activate virtual environment
python -m venv venv
venv\Scripts\activate

# Install dependencies and the project in editable mode
pip install -r requirements.txt
pip install -e .

3. Train the Model

python -m src.train

4. Run the Application Locally

Terminal 1: Start the FastAPI Backend

uvicorn app.main:app --reload

Terminal 2: Start the Streamlit UI

streamlit run ui/interface.py

---

✅ Testing

To run the automated unit tests, first ensure you have installed the testing framework:

pip install pytest

Then, run the following command from the project root:

pytest

---

🎯 Future Goals

• Monitoring: Integrate Prometheus and Grafana for live monitoring of the deployed application's performance and health.
• Advanced Feature Engineering: Experiment with more complex features to further improve model accuracy.
• Hyperparameter Tuning: Implement an automated hyperparameter tuning pipeline to find the optimal model settings.