Institutional RDM policies, repositories and data curation

Data preservation

Four speakers gave compelling presentations. We started with Roxanne Wyns (KUL Libraries-LIBIS), who began by clarifying what we mean by data deposit, retention, and preservation. These concepts—especially the latter two—are often used interchangeably, although they can be used to emphasize different aspects:

Retention primarily refers to the policy dimension: how long data is kept and the process of reappraisal once the initial retention period expires.

Preservation highlights the active care required to ensure continued access to and usability of digital materials for as long as needed. This includes technical actions such as converting data to new formats, transforming metadata schemas, and updating (meta)data content to ensure that it remains understandable and technically usable by research communities in the future.

If data were like lettuce, data retention is about your housemate (or your own acquired common sense) telling you how long to keep it in the fridge and how to decide whether it’s time to throw it out. Preservation, on the other hand, is like removing brown leaves, wrapping it in plastic wrap, or sticking a note to the lettuce that says “mix with tuna for a great salad!”

As active data preservation is one of the requirements of CoreTrustSeal, Roxanne described the process of how KUL's repository RDR became certified between March 2024 and February 2025. RDR is now the first Belgian-based data repository that is CTS-ceritified!

After a brief overview of the actual preservation policy at KUL, Roxanne introduced EOSC-EDEN, which is a European project that seeks to enhance digital preservation strategies at the European and National level, as well as the EOSC-FIDELIS consortium that seeks to enhance the sustainability of EOSC by establishing a network of vibrant and self-sustaining trustworthy digital repositories (see https://eden-fidelis.eu/about-us for more info).

Data curation

Next, after a small break, we invited two experienced data professionals to delve into the practicalities of the current data curation process at their respective institutions. We had the honor of welcoming Kevin Leonard (UGhent, Open Science team) and Marleen Marynissen (KUL Libraries). Both are data curators/metadata librarians with similar years of experience. The main relevant difference between them is that UGhent currently does not have an in-house repository (instead mainly recommending Zenodo), while KUL recommends its in-house repository, RDR.

Kevin opened with the big "why" question: Why do we curate data? Why is data curation important? The answer: because we want to increase the sharing of high-quality data in as FAIR a way as possible.

Building on the empirical findings of three articles, Kevin higlighted, among other things, that perceptions of career benefit, effort, and research impact clearly matter to explain data sharing behavior; and that offering human support to help with infrastructural instruments has a high positive influence on data sharing and reuse. A central message in Kevin’s presentation was therefore that how you communicate curation actions matters as much as the curation actions themselves. Clearly, data curators work not just with the submitted datasets and metadata, but just as much with the humans producing the data. Communication is key.

After showing the actual policy texts regarding data sharing, Kevin described in detail what the data curation process looks like in practice. At UGhent, they use Biblio to register the metadata of published datasets and they currently recommend Zenodo to deposit actual data (and software). Specifically, they ask researchers to submit datasets in the Ghent University Research Data Community on Zenodo. The step-by-step guide is currenlty viewed over 12,000 times since February 2025. With screenshots of real examples, the process became very clear. It is likely that the Zenodo community will continue to exist, once FAIRVault is up and running.

Our third presenter was Marleen Marynissen. Like Kevin, and as requested, Marleen showed the actual policy texts related to RDM, specifically data sharing. KU Leuven has a revised RDM policy since February 2025, encompassing seven principles, with Principles IV and VII being newly added. We started by briefly discussing these principles.

• Principle IV: Everyone who has made a substantial intellectual contribution to a dataset is recognized as author of the dataset
• Principle VII: At the end of research, research data are made available for reuse via a data repository with an open license

At KUL, data curation or review started alongside the launch of RDR in January 2022. All submitted datasets on RDR are reviewed on FAIRness, metadata, documentation, license choice and data protection. Echoing Kevin’s central message, Marleen emphasized that they try to focus on supportive communication, rather than just doing or instructing the actual curation fix. Based on data from the first two years (2022-2023) Marleen showed that researchers are unlikely to make the same mistake twice, when depositing datasets. So, clearly, data curation is actually making researchers better at FAIR data sharing.

Next, Marleen showed us the review dashboard that was launched in May 2023 to keep track of reviews and to avoid double work. Again, it was really informative for us—including my supervisor who gave us a visit—to see how (one aspect of) Open Science is being practised. The review dashboard will only become more relevant now that the revised RDM policy will increase the number of submitted datasets. Last year, in 2024, Marleen and colleagues reviewed close to 200 datasets.

FAIRVault

For our fourth and final speaker, we had the honor of hearing from Jone Paesman (VUB) who presented the (updated) journey of FAIRVault. Some of the new KH members were not yet familiar with the FAIRVault project, so it was relevant to receive an overview of the path so far. At this stage, the project team has decided to use Dataverse to build FAIRVault. As RDR is also based on Dataverse, it may potentially be straightforward to have both repositories communicate with each other. The FAIRVault team is currently working on validating a Minimally Viable Product.

After the three presentations by Kevin, Marleen and Jone, we had a lively Q&A.

In the remainder of the morning, we gave brief updates on three of our Knowledge Forces (i.e., our bottom-up project groups) before we finished the academic part of the day with an “opinion line” game, whereby participants have to position themselves depending on their agreement with each statements. We were only able to do the first two statements (see slides - uploaded on Basecamp) but suffice to say that the statements caused food for thought.

Social activity

In the afternoon, 19 of us enjoyed a lunch paired with a teambuilding game. On a hot day, in a small room, and aided by roleplaying dressed up inspector, our task was to identify the appointed ‘murderer’ at each of the two tables. Despite the topic, this was a nice setting to learn more about each other on a more personal level. I learned about “larping”, geeky SciFi & fantasy book titles and about Lebanese Shawarma.

About our Community Days

The KH Live Days are small conferences we organize 3-4 times every academic year. They are great opportunities to get in touch with fellow data stewards (or professionals in the field of Open Science & RDM more generally). Everyone who is a member of our Basecamp platform is very much welcome.