Docker Data Science

Docker image containing the following Data Science notebooks Jupyter, RStudio and Apache Zeppelin. This docker image is a simplification of KUWADERNO by Paul Ganzon, that you can find at: www.github.com/pganzon/kuwaderno Kuwaderno adds spark and mesos to the configuration that I didn't require. In exchange I have added tensorflow and keras support.

Building the Docker image.

To build the image, clone this repo and run code below.

docker build -t datascience .

If you would not like to build this repository you can use the built image on docker hub.

docker run  -d -p 8787:8787 -p 7777:7777  -p 8888:8888 picarus/datascience

Using the Docker image

Using bridge networking

docker run  -d -p 8787:8787 -p 7777:7777  -p 8888:8888 datascience

or using host networking

docker run  -d --net=host datascience

If you wish to change the port on host networking you can run the following instead.

docker run -d -e PORT0=<rstudio.server.port> -e PORT1=<jupyter.port> -e PORT2=<zeppelin.port> --net=host datascience

After the execution of the code above the following applications should now be accessible.

• Rstudio Server - http://ip.address:8787/
• Jupyter - http://ip.address:7777/
• Apache Zeppelin - http://ip.address:8888/
• Tensorboard - http://ip.address:6006/

Default user and password to access the url are as follows:

• User: admin
• Password: 14mR00t!

You can change both of the user name and password by using the environment variables NB_USER and NB_USER_PASS. For example, if you want the user to be user1 and password1 then run the following:

docker run  -d -e NB_USER=user1 -e NB_USER_PASS=password1 --net=host datascience

Application Versions

Application	Version
Rstudio Server	1.1.447
Jupyter	4.3.0
Apache Zeppelin	0.7.2
Tensorflow

R Packages Installed

As the list is extense and will change often, rather check the file rpackages.R

Python Packages Installed

As the list is extense and will change often, rather check the file requirements.txt

Docker hub

This docker image is also available from:

• https://hub.docker.com/r/picarus/datascience/

Instructions to run from docker

While more comprehensive documentation is put together let's collect some links on how to do it:

• https://docs.docker.com/machine/examples/aws/#step-3-run-docker-commands-on-the-instance

• https://docs.docker.com/machine/reference/env/

• eval "$(docker-machine env default)"

Use from AWS

• (Optional?) Get your Dockerhub image locally: docker pull picarus/datascience

To upload and register the Docker Image on AWS follow the instructions at:

• https://docs.aws.amazon.com/AmazonECS/latest/developerguide/docker-basics.html#use-ecr

(to be added)

Steps to launch docker image