🧠💻
Educating and Training for a FAIR future

Adina Wagner
@AdinaKrik
Psychoinformatics lab,
Institute of Neuroscience and Medicine, Brain & Behavior (INM-7)
Research Center Jülich
ReproNim/INCF fellow



Slides: DOI 10.5281/zenodo.4541323 (Scan the QR code)
Sources: github.com/datalad-handbook/datalad-course

Training for a FAIR future in Neuroscience


Neuroscience strives from interdisciplinarity and collaboration ... but training groups with a diversity of skill sets is difficult

Research data management (RDM)


openaire.eu

... typical difficulties

Lack of formal training
Rarely included in primary/graduate/post-graduate studies. Rather: Learning "on the job" and inheriting project management from the previous person
"Science alone is hard already".
RDM comes on top of everything else, in usually tight, competitive graduate programs or jobs
Too much to know and too little guidance
Rather than a motivational problem, young researchers face the difficulty of finding out which tools exist and are helpful
Late RDM = little benefit
If RDM is only incentivized, required, or tackled as the very last step, researchers can not benefit (fully) from RDM


How can we overcome this?


Personal insights from RDM training

soarperformancegroup.com

RDM overhaul in an institute
User-focused software documentation
    Aim: Researchers turn to RDM because it improves their science and work, not because they are forced to

Training an institute

The DataLad Handbook

Its structure reflects different needs of different stakeholders in science:

    Starting must be easy
  • High-level function/command overviews,
    Installation, Configuration, Cheatsheet
    Tutorials for everyone
  • Narrative-based code-along course
  • Independent on background/skill level,
    suitable for data management novices
    Tutorials for experts
  • Complex functionality and workflows
  • Examples: Computationally reproducible
    analysis on big data
    Overviews for decision makers
  • Step-by-step solutions to common
    data management problems, like
    how to make a reproducible paper

Take Home Messages

  • Complex software tools need accessible documentation
  • Way to adoption of tools or principles via requirements/force/external incentives, or through immediate, personal benefits of good RDM (the latter keeps trainers sane)


    A few Handbook metrics:
  • Three Handbook releases
  • 36 coauthors
  • 500 pages of content
  • dozens of openly shared slides, code lists, and other training materials
  • about 250 unique views per day

Acknowledgements