Jupyterhub @ OCF



With Data Science major going live and producing the first graduates last Fall 2018, the need for readily available data analytics tools has been increasing.

The students enrolled in DATA courses and other connector courses who uses Jupyter/IPython Notebook gets notebook instances through course website. For example, students currently enrolled in DATA 100 can access Jupyter Notebook at https://data100.datahub.berkeley.edu/.

However, students who are interested in Data Science or students who have taken the DS courses in the past are out of luck with accessing cloud instance of Jupyter Notebook. As of January 2019, https://datahub.berkeley.edu/ is available, but there does not seem to have a clearly defined terms of service or policy regarding the maintenance of the JupyterHub instance.

I wanted to gauge OCF’s interest in pursuing and maintaining JupyterHub for our users. Whilst I wanted to bring up the project for discussion, I should note that, this project may be more suitable for a Data Science related club and it to apply for and possibly utilize XCF resources.

Another thing I wanted to note is, without discussing the reason but only to observe, that many students who are pursuing Data Science yet lacking Computer Science background does not know how to run a local instance, hence the need for a managed instance of Jupyter Notebook for DS majors and greater DS community on campus.

JupyterHub installation overview

  • This document Zero to Jupyterhub [GitHub] outlays the minimum steps necessary to get JupyterHub up and running.
  • “4.3.1 Set user memory and CPU guarantees / limits” does not seem to address dynamic allocation of such resources which we will be definitely interested.
  • More to come…