2015 Post-workshop Instructor Debriefing, Round 4

Notes on the meeting can be found here.

Jagiellonian University

Paulina Lach, Leszek Tarkowski, Piotr Banaszkiewicz, and Klemens Noga taught the workshop in Krakow, Feb 21-22, 2015. Some general observations about this workshop were that the etherpad and sticky notes were well received and used throughout the workshop. One of the helpers kept the etherpad up-to-date through the workshop. 41 people were registered at the workshop, of which 36 attended. There was a small amount of attrition of advanced participants on the second day. At this workshop, they collected informal feedback by sticky notes at the end of each section, which the instructors found helpful. They also observed that the workshop participants transitioned from speaking in formal honorifics, normally used in Polish in an academic context, to addressing each other informally, which indicates a good comfort level at the event.

Git:

The instructor covered Git in detail and taught accordingly to Software Carpentry lesson. The participants appreciated the provided handout (http://education.github.com/git-cheat-sheet-education.pdf), a GitHub cheat sheet for Git, although some complained that the Git lesson did not include all of the commands in the handout and would have preferred to see a handout that more closely followed the lesson plan. A quick poll of showed that 14/22 people would have liked to have used pair-githubbing as part of the Git lesson. Generally, the participants felt that the pace was good for Git.

Shell:

The stock "inflammation" lesson material was used. The instructor managed to teach from the beginning through to loops (5 lessons).

Participants were pleased with quizzes at the ending of each lesson, especially the multiple choice questions. Some were disappointed that the lessons covered did not get as far as building shell scripts.

Python:

The instructor started with the "inflammation" material, then used his own examples to place more emphasis on "for" loops and "if" statements (e.g. nested for loops). At the end, the participants were given the challenge of printing prime numbers in the [2,100] range to the screen. Two participants commented that the example was too "mathematical". The participants enjoyed the instructors' funny anecdotes. From quick feedback, almost equal number found the pace "too fast" as did "too slow", which not all the unusual for workshops with a mix of novice and advanced participants.

SQL:

There were some technical issues at the beginning. This topic was taught using the Firefox plugin for SQLite rather than the pure shell. This was the last section taught at the workshop and time was limited, so there was a feeling of being rushed.

Beacon Center at Michigan State University

Jiarong Guo reported on the introductory python workshop held at Michigan State University on January 28-29, 2015. This workshop was structured as three half days, with one each devoted to shell, Python and Git.

Basic Python (data type, variable, list, file stream) was taught using more introductory lessons rather than the standard software carpentry materials. The audience are mostly biologists with no experience in Python or Git. 4/15 had some experience with the Unix shell. The instructors tried to make the materials as introductory as possible. They felt that the Python materials on the Software Carpentry GitHub repo were not basic enough for the audience and that some introduction to basic Python data types was necessary.

Things worked well overall. They spent some time explaining the difference between shell (Bash), Python in the shell, and IPython Notebook, as some participants were confused about these concepts before. Both Python in the shell and IPython Notebook were used during the Python lessons. Hand-on exercises for participants worked really well. When students were not interactive, the TAs helped out by posing questions to the instructors. Some participants expressed that there was not enough discussion prior to the hands-on exercises and that they had no idea how to start. Jirong mentioned that more exercise collections in the repo would be nice.

University of California at Irvine

Jason Moore and Andrea Zonca reported by email about the workshop at UCI Data Science Initiative, held February 21 and 22. The workshop went well.

Some specific notes:

Final feedback from the students is at the bottom of this etherpad
A large percentage of the group had knowledge of one or more programming languages and most knew what variables and loops were. So it was tough to balance going too fast or too slow. We saw this in the pre assessment but still went with the standard material and adjusted as we received feedback. Many people were eager for more intermediate topics so we had a short intro to Pandas that Andrea developed, which went well.
The hosts didn't understand what the sticky notes were for, nor did they notice we needed multiple colors. So maybe that needs to be made clearer in the workshop setup docs. Note from Sheldon: This happens a lot, I noticed they sell post-its at the Dollar Tree, so I will be bringing my own from now on.
There were ~50 students. The first day we had around 6 helpers and it turned out to be more than we needed (probably because the group was more intermediate level than novice at some topics). They used only 3 helpers the second day.
No major technological hurdles. We tried the 5.3 Beta Python and it was a little rough around the edges. Better specify that nobody should use 5.3 for now, it has big issues, like functions being introduced in lesson 4 but used in 2.
It is important to add somewhere in the instructions that the best way to deliver data files to students is with a git pull. It went well at UCI, at UCLA the week before we tried with downloading zip archives and it very time consuming.

The next debriefing meeting will be held on Tuesday, March 10.