Heart Disease Predictor

A machine learning pipeline to predict heart disease presence based on the UCI Cleveland dataset.

Live Demo

Check out the live application here: Heart Disease Predictor

Model Performance

The model was evaluated on a reserved test set (20% of data).

Metric	Score	Description
Accuracy	83%	Overall correctness of predictions.
Recall (Class 1)	0.79	The model correctly identifies 79% of positive heart disease cases.
Precision (Class 1)	0.85	When the model predicts disease, it is correct 85% of the time.

Confusion Matrix

The model successfully identified 22 out of 28 positive cases in the test batch.

Project Structure

• data/: Contains raw and processed data (ignored by git).
• notebooks/: Jupyter notebooks for exploration and prototyping.
• src/: Production-ready Python scripts.
• models/: Serialized models.

Tech Stack 🛠️

This project uses a decoupled architecture with a separate frontend and backend, containerized for portability.

• Machine Learning:
  • scikit-learn: Model training and evaluation.
  • pandas & numpy: Data manipulation and preprocessing.
  • joblib: Model serialization.
• Backend API:
  • FastAPI: High-performance web framework for serving predictions.
  • Uvicorn: ASGI server for production.
• Frontend UI:
  • Streamlit: Interactive web interface for users.
• DevOps & Deployment:
  • Docker: Containerization of services.
  • Docker Compose: Orchestration of multi-container environments.
  • Render: Cloud platform for deployment.

Setup

1. Clone the repo.
2. Run the app on your local machine:
   • Install dependencies: pip install -r requirements.txt
   • Run the API: uvicorn src.app:app --reload
   • Access the Streamlit app: streamlit run src/streamlit_app.py
3. Alternatively, use Docker Compose to run both services: docker-compose up
   • Access the app at http://localhost:8501
   • Access the API at http://localhost:8000

Dataset Description

The dataset used in this project is the UCI Cleveland Heart Disease dataset, which contains various medical attributes related to heart disease. The target variable indicates the presence or absence of heart disease.

Source: UCI Machine Learning Repository

Here's a brief description of the features:

Column	Full Name	Description	Key Values
age	Age	Patient's age in years.
sex	Sex	Biological sex of the patient.	1 = Male 0 = Female
cp	Chest Pain Type	The type of chest pain experienced.	1 = Typical Angina 2 = Atypical Angina 3 = Non-anginal Pain 4 = Asymptomatic
trestbps	Resting Blood Pressure	Blood pressure reading upon admission (mm Hg).
chol	Cholesterol	Serum cholesterol in mg/dl.
fbs	Fasting Blood Sugar	Whether fasting blood sugar > 120 mg/dl.	1 = True 0 = False
restecg	Resting ECG	Resting electrocardiographic results.	0 = Normal 1 = ST-T wave abnormality 2 = Left ventricular hypertrophy
thalach	Max Heart Rate	Maximum heart rate achieved during the test.
exang	Exercise-Induced Angina	Chest pain caused by exercise?	1 = Yes 0 = No
oldpeak	ST Depression	ST depression induced by exercise relative to rest (indicates heart stress).
slope	Slope	The slope of the peak exercise ST segment.	1 = Upsloping 2 = Flat 3 = Downsloping
ca	Major Vessels	Number of major vessels colored by fluoroscopy.	0 to 3
thal	Thalassemia	A blood disorder classification.	3 = Normal 6 = Fixed Defect 7 = Reversable Defect
target	Diagnosis	Presence of heart disease.	0 = No Disease 1 = Disease