Code clubs are initiatives where people meet to improve their coding skills and help each other out. I am a big fan of them as I believe that they can help improving the quality of research. Because as we all know: Better Software, Better Research.

I recently learned more about the code club at the Max Planck Institue of Psychiatry (MPIP) in Munich and would like to share how they organize it in the hopes that it might inspire you. Thanks to the organizers Vera Karlbauer and Jonas Hagenberg for helping me write this post.

CodeClub at MPIP

Target audience: researchers who use code but may have quite varying programming skills (from wet-lab scientists to bioinformaticians).

Goal: increase the quality of their research software and code scripts for better reproducibility and fewer errors.

Structure:

Meetings: 1x per month, ca. 1h

Format:

Short presentations (ca. 10 min) about a relevant topic
Short exercises in small groups

Code Review:

Platform to find a partner
Help with how to perform a code review
Discuss problems and experiences in the group

Implementation:

Ask everyone to present – does not need to be polished
Set realistic goals (only 1h)
Use breakout rooms for exercises when online
Mix experienced with less experienced participants
Collect presentations at one place

Organizers: Two PhD students (currently Vera and Jonas). They recruit a new candidate from the attending students when one of the two PhD students graduates.

Vera presenting the project template that came out of their code club inspired by a template I mentioned to her in a workshop. Made me really happy!

Implementing a code club can be challenging. In their talk, Vera and Jonas mentioned the problems, but they also found some solutions:

Beginners are reluctant to join:
- Communicate and advertise during onboarding
- Low entry barrier and easy first wins (use a project template!)
Wide range of different projects and software.
- Topics that are programming language agnostic but useful for many, e.g. literature review tools.
- Promote a project template for better comparability.
- Best practice template scripts for R and Python (still work in progress).
Often data analysis scripts instead of software development.
- Be pragmatic. E.g. instead of full documentation or github code reviews, code comments may be enough.
Code review is time-consuming and is not considered an important contribution.
- Long-term solution: change culture and recognize achievements.
- Clear code review guide.

Other code clubs

There are many similar initiatives and quite some useful resources. Here is a list of a bunch of them. Please check them out and consider starting join a code club or even starting one ;)