This workshop is designed to introduce users of differing experience levels to new concepts in Python through Jupyter. The problem with traditional lectures (personal opinion) is that I can't reach all of my audience. Either I am too general or too complex and users walk away feeling like they wasted 3-4 hours. Additionally, the resources I host are Linux based which I have been told are intimidating. I have built a JupyterHub at Pitt to give users who are unfamiliar with Linux (majority of them) access to our advanced research computer at Pitt without the barrier of Linux command line, SSH, etc.
At Pitt, I built a JupyterHub server sitting behind the Pitt load balancer supporting single-sign-on. Additionally, we have an Isilon filesystem on the cluster where users files exist. There are some things in the slides which are specific to my instance. You could for example launch these from binder, but I haven't done any testing with this yet.
The workshop was modeled after Jake VanderPlas's Whirlwind Tour of Python.
The key is to be available. The students are going to explore Python by themselves but they are going to have hiccups. I suggest using a system to identify issues. Give every student a blue and red cup. Blue cup is good, red cup is bad. The assistants can then move around and answer questions. If you have any frequently asked questions, address them in front of the class.
I appreciate any and all suggestions. For each workshop, I embed a Google Form at the bottom asking for suggestions (note it is empty because I ask specific questions). Additionally, if you have any suggestions for me, submit an issue here on Github.
- Test binder service with these notebooks for users which don't have accounts at Pitt (and for general usability of the workshop)