Python code and Jupyter notebooks for reference.

Run a notebook server in docker, mounting in this directory with:

docker run --rm -p 8888:8888 \
                -e JUPYTER_ENABLE_LAB=yes \
                -e AWS_ACCESS_KEY_ID=$AWS_ACCESS_KEY_ID \
                -e AWS_SECRET_ACCESS_KEY=$AWS_SECRET_ACCESS_KEY \
                -v "$PWD":/home/jovyan/work \
                --name jupyter \
                jupyter/scipy-notebook:latest

Then run docker logs jupyter to extract the url & token and place these in your browser (something like http://127.0.0.1:8888/?token=33ebd), or use vscode to connect. When you are finished docker stop jupyter.

Alternative environments are listed on https://github.com/jupyter/docker-stacks. I use scipy-notebook when only python is required, or datascience-notebook if R or Julia are required.

Note in the example above we are passing in environment variables from the host. To set an env var on host OSX: echo 'export ENV_VAR=12345' >> ~/.zprofile (or ~/.bash_profile)