❤️ Pakistan Heart Failure Mortality Prediction

---

📑 Project Overview

This repository provides an end-to-end Heart Failure Mortality Prediction pipeline built specifically for Pakistan’s population data.
It demonstrates how data science and machine learning can help predict the risk of heart failure mortality, and how to deploy it for real-world use.

---

📂 Contents

• 📒 Jupyter Notebook

  • Detailed EDA, preprocessing, feature engineering, and model training.
  • Visualizations and performance evaluation.

• ⚙️ Streamlit App (app.py)

  • A simple, interactive web app to test predictions using the trained model.
  • Easy to run locally or deploy online.

• 🐍 Upcoming Python Project

  • The next version will include a FastAPI implementation for robust API deployment.
  • Supports integration with modern frontend or mobile apps.

• 🐳 Planned Docker Deployment

  • A Dockerized version for easy deployment to cloud servers or containers.

---

📊 Dataset

This dataset contains comprehensive cardiac patient information for heart failure mortality prediction from Pakistani patients, collected by the Computer-Human Interaction and Social Experience Lab (CHISEL) at LUMS University.

Source: Open Data Pakistan - Heart Failure Prediction Dataset URL: https://opendata.com.pk/dataset/the-heart-failure-prediction-pakistan Total Records: 368 patients Total Features: 60 (59 features + 1 target variable)

Disclaimer: The dataset is limited in scope and scale. It is meant for research and educational demonstration only — not for real clinical diagnosis.

---

🧮 Data Analysis & Model

• Exploratory Data Analysis (EDA)

  • Detect outliers, missing values, and correlations.
  • Visualize distributions and relationships.

• Preprocessing & Feature Engineering

  • Categorical encoding, scaling, and handling missing data.
  • Feature selection for best model performance.

• Machine Learning Model

  • Random Forest Classifier — chosen for its balance of performance and interpretability.
  • Evaluated using metrics like Accuracy, ROC-AUC, Confusion Matrix, and Feature Importance.

---

🚀 Deployment

✅ Current Status:

• Streamlit App: Users can input patient data through a simple UI and get instant predictions.

📌 Planned:

• FastAPI Backend: Serve the model as a production-grade REST API.
• Docker: Build and deploy the API and app in containers for easy scaling.

---

📷 User Interface

---

⚙️ How to Run

1. Clone this repo:

   git clone https://github.com/ahsan-javed-ds/pakistan-heart-failure-mortality-prediction.git
   cd pakistan-heart-failure-mortality-prediction

   pip install -r requirements.txt

   streamlit run app.py

OR

2. Simply run the attached Colab Notebook: Pakistan Heart Failure Mortality Prediction Notebook

✅ Contribution & Impact

This project highlights how AI solutions can support healthcare in Pakistan by:

• Providing early risk prediction tools for hospitals and clinics.

• Encouraging data-driven awareness of key mortality factors.

• Demonstrating how local data can be used to build local solutions.

⚠️ Limitations

• The dataset is small and may not cover all demographic and medical variations.

• Predictions are not medically certified — they are for educational and research purposes only.

• For real-world use, larger, validated datasets and clinical testing would be required.

🤝 Contributions

All contributions are welcome!

📌 Ideas for improving the notebook, app UI, expanding the dataset, or adding new deployment methods (FastAPI, Docker) are highly appreciated.

Please open an issue or submit a pull request to collaborate.

📜 License & Disclaimer

This project is open for educational and non-commercial research use only. Always consult qualified healthcare professionals for any real medical decisions.

📧 Contact

Author: Ahsan Javed

• Email: [email protected]
• GitHub: https://github.com/ahsan-javed-ds
• LinkedIn: https://www.linkedin.com/in/ahsan-javed17/