Simple template demonstrating how to set up a docker container for reproducible data science proof-of-concept projects.

• Clone this repository and entire directory.

git clone [email protected]:eugeneyan/poc-docker-template.git
cd poc-docker-template

• Build the docker image (make sure you have docker running).

docker build --tag poc-docker-template .

• Run the container and run.sh.

docker run --rm -v $(PWD):/poc-docker-template --name tmp -t poc-docker-template /bin/bash run.sh

• (OPTIONAL) Clean up the container.

docker rmi poc-docker-template

• Check Jupyter notebooks.
  • You can view the visualizations in 2b-data-visualization.
  • Machine learning results are in 4-train-model with timestamps on the latest logs.

cd notebooks
jupyter notebook

To use it with your project, update the following files:

• data/: Replace with your datasets.
• notebooks/: Replace with your notebooks.
• requirements.txt: Replace with the packages and versions you use, or use poetry.
• run.sh: Update to run your Jupyter notebooks in order.