Public | Automated Build

Last pushed: a year ago
Short Description
A customized jupyter/datascience-notebook.
Full Description

Forespell data science stack

This repo contains (1) a Dockerfile with Jupyter notebook data science stack whose image is built by Docker Hub when master is updated, and (2) a script to launch the Docker container on a powerful AWS EC2 spot instance on with an attached EBS volume for data persistence.

One-time setup

  1. Install the AWS CLI with pip install awscli. Install your AWS credentials under ~/.aws/credentials.
  2. Create an EFS in the region eu-west-1 in the AWS console and give assign it the Name: docker-notebook-fs.
  3. Add an inbound rule to the default security group in the AWS console to allow traffic from port 443.
  4. Set the environment variable CLOUDFLARE_API_KEY. On Linux, use ~/.bashrc, on macOS use ~/.bash_profile and on Windows use setx.
  5. Use IPython.lib.passwd() to generate a hash of the notebook password, and update the NOTEBOOK_PASSWORD variable with it.

Starting and stopping the notebook server

To start the notebook server, simply run on Linux and macOS, or notebook.bat on Windows. Then go to To add an exception for the self-signed certificate, see

To stop the notebook server, simply run docker-machine rm supercomputer.

Docker Pull Command
Source Repository