💓 Heart Failure Predictor

A machine learning–powered web application that predicts the risk of heart failure using clinical data. Built during DevTown’s Predictive Modelling Bootcamp, this project combines model training, Flask-based deployment, and a custom web interface.

---

📌 Project Overview

• 🔍 Objective: Predict whether a patient is at high or low risk of heart failure
• 📊 Model Accuracy: Achieved over 80% on test data
• 🛠️ Technologies: Python, Scikit-learn, Flask, HTML/CSS
• 📁 Dataset: Heart Failure Clinical Records Dataset

---

📂 Project Structure

heart-failure-predictor/ ├── app.py # Flask backend ├── model_training.ipynb # Model training notebook ├── model.pkl # Trained model file ├── scaler.pkl # Scaler used for input normalization ├── heart_failure_clinical_records_dataset.csv # Dataset ├── templates/ │ └── index.html # Web form frontend ├── static/ │ └── style.css # Custom CSS styling

---

🧠 How It Works

1. User inputs 12 clinical features (age, BP, sodium, etc.)
2. Inputs are scaled and passed to the trained ML model
3. Model predicts the probability of heart failure
4. Result is displayed with an intuitive message & emoji

---

💡 Sample Test Inputs

Try this in the form to test a high-risk case:

Age: 75 Anaemia: 1 Creatinine Phosphokinase: 500 Diabetes: 1 Ejection Fraction: 25 High Blood Pressure: 1 Platelets: 200000 Serum Creatinine: 2.1 Serum Sodium: 125 Sex: 1 Smoking: 1 Time: 4

yaml Copy Edit

👉 Expected Output: 💔 High Risk of Heart Failure

Form Page Prediction Output

🏁 Built With 🐍 Python

⚙️ Scikit-learn

🌐 Flask

💅 HTML + CSS

🙏 Acknowledgments

DevTown Predictive Modelling Bootcamp Heart Failure Dataset from Kaggle