Installation via Docker

We're building this container on top of Debian 8 Docker container and Anaconda Docker container.

Installation of Docker Daemon

Download and install the Docker Daemon from the Docker Website for your OS.

Proxy Configuration in the Docker Daemon

Docker will download all the necessary files for the build process from the internet so it might be necessary to configure a proxy. You can switch to a manual proxy configuration in the Docker Daemon Settings under Proxies.

Build Docker Container locally

Now you need to build the docker container. Navigate in the terminal into the repository directory and then into the docker directory machine-learning-workflow/docker.

Proxy Configuration in the Dockerfile

In case you need a proxy and you have set it correctly in the Docker Daemon, you need to set it in the Docker file as well. Uncomment the four corresponding lines under the both # for proxy usage headlines (on the top and bottom of the file) and add your proxy address in the first two lines. The last two lines remain empty so there is no need for a change.

Execute the Build Command

Now execute in the machine-learning-workflow/docker directory the following command.

docker build . -f Dockerfile -t ml_workshop_container

Run Docker Container

After the build process was successfully done you can run the container.

docker run -it -p 8888:8888 -p -d -v <absolute/path/to>/machine-learning-workflow/notebooks:/notebooks ml_workshop_container

Use Image from DockerHub

docker run -it -p 8888:8888 -p -d -v <absolute/path/to>/machine-learning-workflow/notebooks:/notebooks datascienceworkshop/machine-learning-workflow

We're using the following parameters:

• -p 8888:8888 to export Jupyter Web interface
• -d to run Docker container in background
• -v notebooks:/notebooks to mount notebooks folder inside Docker container

Accessing and Running the Notebooks

The notebooks are now available in your browser under the following url:

http://localhost:8888