Follow-up on PVSC50 pvlib workshop

[AdamRJensen]

@mikofski @shirubana I'm curious as to how the workshop went? Any feedback for future workshops that would be useful to share? 😃

[mikofski]

Huge kudos to @shirubana for making this event happen and for introducing the problem based learning approach. The PVSC tutorial structure was about 3 hours long with a break, and there were about 25 attendees. We split the time using the first half to go through tutorials 0, 1 & 2 as in previous events, but used the second for the problem based learning approach which we did in groups of 4, so there were 6 groups. Silvana had one of the blind modeling challenges prepared which we reused for the challenge. The problem based learning was a success. Although it’s scary, I would recommend it every time from now on. The attendees dove headfirst into the challenge with gusto! Also Colab worked great. Some attendees participated from their mobile devices.

Issues:

• Timezones, especially native vs. localized, was still an issue.
• Jupyter, cell order, so confusing, rerunning, not running cells, what has been defined, parsing errors took time to trouble shoot.

Suggestions:

• I believe we should start thinking about an advanced track.
• Not specific to this tutorial, but a lot of people have come up to me at the conference with sideline conda environment issues, coming from MATLAB this managing environments, editors, packages, and jupyter kernels is a real issue, IDK what the answer is, but it’s a blocker for a lot of folks who want to advance past the tutorial
• Other packages need to be addressed in other tutorials: rdtools, pvanalytics, pvmismatch, etc.
• We need to spend EVEN MORE TIME teaching folks to use documentation, some folks have asked me outside of the tutorial to explain things that are documented

[shirubana]

We did 2-3 tutorials (0 and 1 more heavily, 2 after break and fast paced but enough), and then had them work as groups in an empty jupyter journal to do the PBL 2 (PVPMC setup 1 from 2021 round robin). It was absolutely awesome. Other things that worked out:
1- Slightly longer introductory session of ourselves and the motivation of the tutorial, as well of each of them (25!) expectation to build rapport. This took about 30 mins.
2- After every tutorial, asking ‘5 things you learned’ from the audience (5 things for all of them, not individually)
3- Having them stand up with the slides from PVPMC 2021 and reorganize themselves into groups by python proficiency

Things that I would suggest:
-I know the conflict of the 30 minute sun shift is dear to our heart, but it really looses people. This should only be in the advanced track
-Make an Advanced track

Advanced track suggested topic coverage:
-Raw(ish) set provided of field data; load it, format it and do QA on it.
-Right-left labeling and sun position convention
-Uncertainty/error calculations and/or sensitivity analysis (this can include adding noise to the data, or sorting of results for uncertainty, or talking about sensors, a small monte carlo simulation, or a range of values to introduce to PVLib showing how to make a bit of sense of the data; I seem to remember also another python packages that can be used on top of pvlib to handle this…).