Data stewardship
Résumé of the "Data Stewardship goes Germany" conference in Dresden, 2023.
Data Stewardship - what does that mean?
On September 25th – 26th the Data Stewards of the SFB135 and TAM attended the “Data Stewardshop goes Germany”-Workshop at the SLUB Dresden. Goal of the workshop was to bring together Data Stewards from different German Universities and present and discuss each other’s roles and tasks.
The kick-off was made by Jens Dierkens who published the DataStew Report. In this project and report the team of Jens Dierkens discuss and measure the concept, education, and expertise of Data Stewards as well as the needs of researchers for a Data Steward for their research work. Very briefly, the report shows that there is no such thing as “the” Data Steward. Tasks of Data Stewards highly depend on their environment, e.g., if their responsibilities concern a certain CRC, lab, or the whole university. The report also gives an overview of competencies, profiles, and tasks of Data Stewards and gives recommendations for institutions (see the talk about the report here).
During the workshop the organizers set up so called BarCamps in which we discussed in small groups the objectives of the report. We discussed the roles and tasks, centralized vs. domain-specific responsibilities and knowledge, service or science focused work, the scope of Data Stewardship, and training of Data Stewards (results of all BarCamps in this GoogleDrive).
What became clear is that:
- Tasks of Data Stewards depend on the kind of employment (university, CRC, lab; not actually a Data Steward but responsible for data management in the lab) and on the field (in some fields there are no to very few established standards on e.g., metadata)
- However, the three main tasks we almost all had in common was training, consulting and creating workflows/structure
- Basic RDM knowledge needs to be part of researchers’ education
- Most Data Stewards wish for more support from PIs or department principals
- Data Stewards need to have domain specific knowledge (i.e., prior or present research experience in the field) to be able to understand the needs of the researchers
Furthermore, we presented our Data Hub workflow which includes BIDS, Git, GitLab, DataLad, and GIN (see our presentation). We realized that with BIDS we can consider the field of psychology and neuroscience as advanced regarding a widely adapted common folder structure and metadata specification. On top of that, there’s lots of software available to convert data automatically in the specified structure and extract the desired metadata. In other fields, Data Stewards have to think of all this themselves. Thanks to BIDS, we as Data Stewards can focus on providing you with an almost-automated workflow of managing your data using the tools mentioned above.
Lastly, one discussed topic was also the incentives of the researchers to do RDM. Besides advantages such as easy collaboration, reproducibility, credibility and so on, we would like to bring in another incentive specific to our workflow provided by the Data Hub:
All the tools we utilize become more and more popular in the field of neuroscience. When you look at job advertisements for e.g., PostDoc positions, especially outside of Germany, you now see it quite often that basic knowledge about at least BIDS and Git is listed as a requirement. A lot of labs do their daily work on GitHub. In the fields where you have to handle big binary data such as imaging, DataLad is going to gain more attention, as well.
Seeing it from this point: Do yourself a favor and make use of our services and support to prepare for your future in neuroscience research :-)