Skip to content

Infrastructure & Tools

Decide how your students will run notebooks. The materials on this site are platform-independent Jupyter notebooks, so several options work.

Need a hub without adopting curriculum?

Cal-ICOR and Cloudbank can provision computing infrastructure independently - you don't need to adopt a course or module to get a JupyterHub. For Cal-ICOR, email calicor@berkeley.edu or fill out the interest form. For Cloudbank, email Sean Morris.

Notebook delivery options

Campus JupyterHub

If your institution already runs a JupyterHub, students sign in with campus single sign-on and notebooks open directly. You'll use the Canvas JupyterHub rewriter to point assignment links at your hub.

Cal-ICOR

Cal-ICOR (California Interactive Computing Open Resource) provides cloud-hosted computing infrastructure for California public colleges and universities. If your institution is in California and doesn't already run its own hub, Cal-ICOR can provision one configured to your campus sign-on. Provisioning is scheduled, so start early. Email calicor@berkeley.edu or fill out the interest form.

Cloudbank

Cloudbank is an NSF-funded program that provisions cloud computing infrastructure, including JupyterHubs, for institutions nationwide. If your institution is outside California or needs broader cloud resources, Cloudbank is a good fit. Provisioning is scheduled, so start early.

Other options

For piloting or smaller deployments, the notebooks also run on Google Colab or Binder. These are good temporary options while a hub is being prepared. Longer term, other platforms such as NRP, Jetstream, Vocareum, and Codespaces may also be supported.

What you need

  • A GitHub account (free): create one here
  • A way for students to run notebooks (one of the delivery modes above)
  • An LMS such as Canvas if you want to use our cartridge (the notebooks also work with other LMSs)
  • A commitment not to distribute solutions publicly

Privacy and accessibility documentation

Your institution may need documentation related to privacy (HECVAT) and accessibility (VPAT). The student information required is minimal (an email address). Accessibility is an ongoing effort across UC Berkeley and the Jupyter community; ask us for the current HECVAT and accessibility statements when you adopt.

A deeper technical reference

For a thorough overview of the infrastructure needed to run a data science course, see The Data Science Educator's Guide to Technical Infrastructure.