By Audrey Knauf, Associate Professor, Crem (University of Lorraine) – Open Science Ambassador, University of Lorraine – Deputy Director Data & Corpus – the journal of data in SSH – Co-leader of the ANR SoSHS project
As part of Love Data Week at the University of Lorraine, the study day dedicated to research data brought together speakers on March 19th to share experiences, surveys, methods, tools, and reflections on a central issue for us researchers: training in research data.
One of the major contributions of this day is the reminder that data are not simply a “side note” of research. On the contrary, they play a central role in our practices, at the heart of our most ordinary actions, sometimes invisible, while engaging in something fundamental: our way of producing knowledge, making it intelligible, transmitting it, and ultimately responding to it. They contribute to evidence, method, scientific discussion, work memory, and the potential for circulation.
The starting point of this day was clear: reflecting on skills, learning formats, their integration into teaching paths, and their adaptation to disciplinary specificities, making training a decisive lever. The transformation of practices requires a true practical data culture beyond mere adherence to open science. Among the training offerings presented are ADOC Lorraine’s data workshop, designed throughout the data life cycle, DoRANum as a self-training platform, and the FLSO project, training doctoral students in open science through peer transmission.
At the core of the exchanges, the Open Science Barometer indicates a lower rate of open access in SSH, without implying a lack of open practices. The insight from the SoSHS project invites nuanced perspectives: it highlights existing, discreet, less standardized, and less visible practices in the indicators. This leads to transcending the distinction between disciplines “lagging behind” and “ahead.”
In other words, data remind us that research always takes place in concrete conditions with objects, methods, audiences, and legal, ethical, and scientific responsibilities, an essential nuance that applies to disciplines and data types alike. The afternoon examples illustrate this: the ArchiMed platform in clinical research involves structuring, traceability, and security issues, while the Les Vocaux corpus, for language data, emphasizes openness, uses, and participative frameworks. Despite their differences, a common requirement emerges: to consider data in their context of production, circulation, and reuse.
This is why this issue resonates so much with the research community: it entails responsibility because openness cannot be considered without regard to individuals, rights, fields, and production contexts. It goes beyond a property-centered approach: with the principle of open access by default for public data, the question becomes about their nature, legal regime, and specific conditions for openness. It also challenges the very organization of scientific work, showing that research can no longer be seen as an individual and isolated activity.
Research data is a collective issue for the academic ecosystem. Open practices remain more limited without data engineers, who are central socio-technical mediators (cited by Justine Richard) in reshaping the roles brought by open science. Behind data sharing and FAIR principles lies a considerable, often invisible but essential, effort requiring support, recognition, training, resources, and dialogue between stakeholders.
This day shed light on a convergence of worlds that have sometimes worked side by side without considering each other (researchers, librarians, support staff, legal experts, data engineers, IST specialists, infrastructures, and platforms), all contributing, in their own way, to the quality of what we produce and transmit. In conclusion, it was suggested that these discussions continue in laboratories, projects, doctoral programs, fuel team discussions, and that the idea of Nicolas Fressengeas “the main obstacle is us” should belong to the past. Taking care of research data is also taking care of research itself.
