MLOP-Project: Linear Regression Model for Predicting Math Scores

Overview

This project demonstrates a Linear Regression model used to predict students' math scores based on various features. The project includes exploratory data analysis (EDA), model training, and deployment on both AWS and Azure platforms using Flask.

Project Structure

MLOP-Project/
├── Notebooks/
│   └── EDA/                       # Contains Jupyter notebooks for Exploratory Data Analysis
├── Template/
│   └── index.html                 # HTML file for the web application
├── application.py                 # Flask application to serve the model
├── setup.py                       # Script to set up the environment and install requirements
├── requirements.txt               # List of dependencies required for the project
├── Src/
│   ├── Component/
│   │   ├── data_ingestion.py      # Script for data ingestion
│   │   ├── data_transformation.py # Script for data transformation
│   └── Pipeline/
│       ├── training_pipeline.py   # Script to handle the training pipeline
│       ├── prediction_pipeline.py # Script to handle the prediction pipeline
├── .github/
│   └── workflows/
│       └── azure_webapp.yml       # GitHub Actions workflow for deploying to Azure Web Apps
└── .ebextensions/
    └── option_settings.config     # Configuration for deploying to AWS Elastic Beanstalk

Getting Started

Prerequisites

• Python 3.x

• Required Python libraries can be installed using:

  pip install -r requirements.txt

Running the Project

Clone the Repository:

git clone https://github.com/YourUsername/MLOP-Project.git
cd MLOP-Project

Set Up the Environment:

python setup.py install

Run the Flask Application:

python application.py

The application will be available at http://localhost:5000/.

Project Components

• Exploratory Data Analysis (EDA):

  Jupyter notebooks located in the Notebooks/EDA/ folder provide insights into the dataset used for model training.

• Web Application:

  The web interface for the model is served using Flask. The HTML template for the app is located in the Template/ folder.

• Source Code:

  • Src/Component/: Contains scripts for data ingestion and transformation.
  • Src/Pipeline/: Contains scripts to manage the training and prediction pipelines.

Deployment

AWS Elastic Beanstalk

The project includes an .ebextensions/ folder containing configuration settings for deploying the Flask application on AWS Elastic Beanstalk.

To deploy, ensure you have the AWS CLI set up and run:

eb init
eb create

This will initialize and create an Elastic Beanstalk environment, deploying the application automatically.

Azure Web Apps

The .github/workflows/ folder contains a GitHub Actions workflow for deploying the application to Azure Web Apps.

The deployment can be triggered by pushing to the main branch. The workflow defined in azure_webapp.yml automates the deployment process.

More information on the Azure Web Apps deployment action can be found here: Azure Web Apps Deploy.

License

This project is licensed under the MIT License. See the LICENSE file for more details.