Project Overview

This project containerizes a machine learning Python flask app using Docker and deploys it on a Kubernetes cluster. The app predicts housing prices in Boston according to several features, such as average rooms in a home and data about highway access, teacher-to-pupil ratios, and so on. The data was initially taken from Kaggle.

---

Running app.py

1. Standalone: python app.py
2. Run in Docker: ./run_docker.sh
3. Run in Kubernetes: ./run_kubernetes.sh

Making predictions

In a separate terminal, run: ./make_prediction.sh

---

Project Structure

The project includes 4 directories:

• model_data, containing the machine learning models
• output_txt_files, containing the output of running a prediction on Docker and on Kubernetes
• .circleci, containing the configuration file to implement CI/CD through CircleCI
• .git, containing the git files

It also includes the following files:

• app.py, the main Python flask app
• Dockerfile, to define a Docker container
• LICENSE, containing info on the license
• Makefile, to build the project
• make_prediction.sh, to run predictions
• README.md, this file
• requirements.txt, containing the dependencies that are installed through the Makefile
• run_docker.sh, to launch a Docker container
• run_kubernetes.sh, to deploy containers on a Kubernetes cluster
• upload_docker.sh, to upload a Docker image to Docker Hub
• .gitignore, to list files that Git should ignore

---

Authors

• Antonella Bernobich Dean - aberdean

License

This project is licensed under the MIT License - see the LICENSE file for details.