Reproducible Data Science with Docker

• As presented at https://www.meetup.com/Docker-Dublin/events/247314882/
• ASCIINEMA DEMO: https://asciinema.org/a/165285
• Commandline Interactive Tutorial / Demo (as recorded above) can be found in ./docker-demo. Entry point is menu.sh. Requires dialog, figlet and docker (can be installed using docker-demo/tools/install-docker-ubuntu.sh)

Docker image:

• The final image built by this demo can be found at: https://hub.docker.com/r/cmihai/jupyter/

Docker Compose file and scripts

• Source can be found here: https://github.com/crivetimihai/docker-data-science

How to run the image:

• This image has been published as cmihai/jupyter. Just do a:

docker pull
docker run --init --detach \
    --name jupyter --hostname jupyter \
    -p 8888:8888 \
    --volume ${HOME}/notebooks:/home/jupyter/notebooks:rw \
    --volume ${HOME}/.jupyter:/home/jupyter/.jupyter \
    cmihai/jupyter

Alternative Images:

https://github.com/jupyter/docker-stacks/tree/master/datascience-notebook is quite popular.. you can use that instead.

docker run -it --rm -p 8888:8888 jupyter/datascience-notebook

Reference and where to learn more:

• Learn Docker here: https://www.katacoda.com/courses/docker
• Learn Shell here: https://www.hackerrank.com/domains/shell/bash