Heart Disease Prediction Web App

Flask app that predicts the risk of heart disease based on a Gradient-boosted Trees ML model, and shows the confidence in the prediction as well as the factors behind the prediction (explainability).

Data collected from this dataset. EDA and model development code available in this repo.

You can access a live website running the application here.

Getting Started

Dependencies

Docker

Linux: To install Docker on Linux, follow the instructions for your specific distribution on the Docker website.

Windows: If you're using Windows, you can install Docker Desktop by downloading it from the Docker Desktop for Windows page.

Docker Compose

Linux: To install Docker Compose on Linux, follow the instructions on the Docker Compose website.

Windows: Docker Desktop for Windows includes Docker Compose. You don't need to install it separately.

Installing

To install this application, follow these steps:

1. Clone this repository to your local machine:

   git clone https://github.com/leo-cb/HeartDiseasePrediction_WebApp.git

2. Set environment variables with MongoDB username and password:

   Linux:

   export MONGODB_USERNAME=<your_mongodb_username>
   export MONGODB_PASSWORD=<your_mongodb_password>

   Windows:

   set MONGODB_USERNAME=<your_mongodb_username>
   set MONGODB_PASSWORD=<your_mongodb_password>

3. Start the docker-compose service:

   docker-compose up -d

Executing program

Once the docker-compose is up, you can access the application by opening a web browser and navigating to http://localhost:5000. From there, you can enter patient data and receive a prediction of their likelihood of developing heart disease, as well as the confidence in that prediction and the each factor weight behind it.

Description

Overview

The application serves several key functions:

1. Model Loading: It loads a GBT binary classification model from a pickle file.
2. API Endpoint: The application creates an endpoint at /predict, which handles API requests. To access this endpoint, a valid API key (stored in a MongoDB collection) is required. It provides predictions (0 or 1), confidence percentages, and SHAP values for each variable used in the model.
3. API Key Management: It includes library files to create API keys and insert the hashed key into the MongoDB collection. It also provides a function to check if a given API key exists.
4. Unit Tests: The application includes unit tests located in the /tests/ directory.
5. Web Interface: It generates a web page called index.html, allowing users to input variable values and submit them. The output is displayed in results.html.

Continuous integration/Continuous deployment

CI workflow

The repository has a CI workflow in .github/workflows named python-webapp.yml that runs on every push request on both main and dev branches. It runs linting and runs the tests in /tests/.

CD workflow

The repository has a CD workflow in .github/workflows named deploy.yml that runs on a successful CI workflow run on the main branch. It deploys the application to a Debian-based host by ssh access, and runs the docker containers.