Skip to content
Dockerfiles for various machines used in classes
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
class-tools-infrastructure
.travis.yml
Makefile
README.md

README.md

Using Docker to deploy a data science environment

This repository contains our setup and configuration files for setting up a data science environment for the students at NYU/Stern. (Jupyter running Python and R, plus MySQL).

Instructions for students

  • As the simplest, and default option for NYU students, we offer a JupyterHub server, running at https://jupyterhub.stern.nyu.edu. You need to be within the NYU network to connect, or use the NYU VPN. Login using your NYU id as a login (e.g., abc123, not your email) and use your NYU password to log in.

We also a few more options for students (or classes) that want more flexibility:

Features of JupyterHub installation at NYU

  • We provide directly support for Python 3 and R kernels. If you want additional kernels let us know.

  • By default each student that logs in gets access to a (containerized) 6CPU / 53Gb Ubuntu machine running Jupyter.

  • Auto-sharing notes through Github: If you have class material stored on Github, we can add your class on JupyterHub, and all your notes will appear automatically under the /notes folder, when students login: We just need to add the URL of your Github repository in the courses.yaml file.

  • Resolving conflicts: If you update your notes on Github, and your students have modified the earlier version of the notes, when we fetch the latest version of the notes, we also make a backup of the earlier file that was modified by the student. The clone_nbs.py script is the one that clones the repositories, makes backups of the conflicting notebooks, etc.

  • Sharing data files: If you want to share data files that are too big to be stored on Github, we can upload the files on JupyterHub, or we can even setup a mounted folder on JupyterHub, also mounted on your local computer, which you can use to share files with the students.

  • Adding new software/libraries: If you need an additional Python or R library (or some other Unix software) to be available on JupyterHub, we can add it to JupyterHub by simply modifying the configuration Dockerfile. In the future, we plan on supporting Binder, which will allow faculty to configure their software environment directly from their Github repositories.

  • Support for nbgrader: We have experimental support for nbgrader built on JupyterHub. Let us know if you want to use it.

  • Support for JupyterLab: We have installed JupyterLab as well. If you prefer the JupyterLab interface, just replace the /tree in the URL with /lab.

You can’t perform that action at this time.