Within Inria, the organisation around data is structured around the research data hub of the IES (Information et Édition Scientifiques) Service. This cluster supports scientists in the drafting and proofreading of data management plans and the opening of data via the repository and archiving on the Inria institutional space of Recherche Data Gouv. This cluster is supported by the members of a national “research data” unit, led by the Directorate of the IES service and which brings together representatives of the various Inria services concerned to provide the best possible support to project promoters (DPO, IT Systems Department, Defense and Security Official, Coerle, archivist, etc.). The IES data hub also leads the network of research ‘data referents’ composed of a scientific representative and an IES correspondent for each research centre in relation to the national Recherche Data Gouv ecosystem and the European projects in which Inria is currently involved.

This organization, which combines an operational and scientific approach, promotes a shared vision of news and data-related projects.

Interview with the data hub team

Is the space backed by a competence centre?

The Inria open data repository is linked to the local competence centres and according to the implementations of the Inria centres: ABDo (Atelier Bordelais des données de recherche), ADOC (Accompagner aux DOnnées les Cherchers et chercheurs en Lorraine), ARDoISE (Atelier Rennais de la Donnée), CDGA (Cellule Codes Données Grenoble Alpes), DATALystE (Atelier de la données Université de Lyon), DatASaclay (Dat’Atelier de l’Université Paris-Saclay), DATAZUR (Atelier de la données Université Côte d’Azur), EcoDoR (vers un écosystème des données de la recherche sur le site de Montpellier), LORD (Lille Open Research Data).

Data referent scientists and members of the IES service contribute to the activities of Geographic Competence Centres and national ecosystem working groups (e.g. the Code and Software WG of the CDGA office).

Which disciplines are primarily represented by this data?

Topics most represented among the submitted datasets are those related to digital sciences and particularly related to medicine, health, life sciences, agriculture, earth sciences and the environment.

What is the process for submitting data? For creating a collection?

Anyone connected to the French ‘Renater’ identity federation and having created an account on Recherche Data Gouv can deposit datasets in the institutional space. The deposit at the Inria repository is reserved for Inria’s member project teams. The collections are created at the request of scientists by the data center that accompanies them for the implementation. Each user-specified request is handled by the same hub and may lead to changes in service provision.

What curation is carried out on the platform?

The curation is carried out before the validation and uploading of the dataset. As mentioned above, the curation is carried out by the members of Inria’s data hub. The curation is based on the grid produced by the Recherche Data Gouv team.

In the responses, recommendations are regularly proposed on the licensing of the dataset, the software described and particular attention is paid to the archiving of code and software on Software Heritage and links to other research products such as publications.

Is there any integration with other repositories?

Scientists' datasets are deposited primarily in thematic data repositories. Datasets deposited in other institutional spaces of Recherche Data Gouv are referenced in the Inria collection (by the functionality “Link dataset”).

Do you have a model dataset or one you would recommend as an example of a ‘clean’ deposit?

A good example is the 4DHumanPercept’ dataset:

REKIK, Rim; WUHRER, Stefanie; HOYET, Ludovic; ZIBREK, Katja; OLIVIER, Anne-Hélène, 2025, “4DHumanPercept Dataset”, DOI: 10.57745/NZHDFY

This dataset serves as a baseline for studies related to immersive technologies (virtual reality, creation of avatars). The dataset was created for the study ‘Quality assessment of 3D human animation: Subjective and objective evaluation’ and allows researchers to train and test algorithms so that they are able to judge the quality of a 3D/4D animation as a human eye would.

Find out more

• Institutional desk: data.ies[at]inria.fr
• URL for the institutional space: https://entrepot.recherche.data.gouv.fr/dataverse/inria
• Open science at Inria: https://www.inria.fr/en/open-science-inria-role

Natural language search, independent of the language of description

The catalogue’s search engine accepts queries phrased as questions, for example, ‘data on drought in the Mediterranean region since 2010’. It relies on the meaning of the query rather than an exact match of terms, which makes it possible to find datasets even when their descriptions are written in another language.

The results can then be refined by discipline, organisation, date or author.

An assistant capable of searching the catalogue

This development is based on a technical standard, the MCP (Model Context Protocol), which enables an AI assistant to query the catalogue directly and utilise the descriptions of the datasets. The assistant can retain up to 200 records during a single session, which opens up several possible uses:

- identify the teams and key contacts for a particular topic, with a view to collaboration;

 - compare the summaries and methodologies of several datasets, without having to consult them one by one;

 - identify cross-disciplinary links as the discussion progresses.

An open testing phase

A dedicated chatbot is available to try out the system, with no installation required. Anyone interested is invited to contact the team via the Recherche Data Gouv portal to join the test phase: contact the team via the portal (select: ‘Provide feedback’ or ‘Get help with the catalogue’) to join the test phase.

Find out more: Recherche Data Gouv catalogue page

This ressource is the result of a collaboration between competence centers of Recherche Data Gouv. It guides you step by step through the selection and long-term preservation of your data, in collaboration with your archivists.

Record management or permanent preservation of research data is an essential step in ensuring that data remain accessible for as long as necessary. The guide sets out the main criteria to select which data should be permanently preserved, taking into account their scientific value, documentation, quality, and compliance with applicable standards and guidelines. It also addresses the legal and ethical considerations associated with their preservation.

Aimed at a broad audience—including researchers, research teams, laboratory managers and data management professionals—the guide is above all a practical tool that can be readily incorporated into day-to-day practices, particularly when drafting data management plans.

Direct access to the resource on DoRANum

Share it with your teams to promote a common approach to data selection, harmonise practices and better distinguish between data that should be retained and data that may be disposed of, in compliance with regulatory requirements.

Structured in ambitions, challenges, strategic objectives and action plan, the French ‘Stratégie de science ouverte’ (SSO) devotes an important part to research data. On this aspect, it recalled that ‘data sharing is not yet the norm’ (p. 31) but that it is increasing by about 2% per year, with the latest key figures from the french Open Science Barometer showing that 27% of publications have shared their data.

The citation of scientific sources is also recalled as essential, in a context of development of generative artificial intelligences. GenAI is based on data available online, so "the free dissemination of reliable and validated research results contributes to enriching the quality of information" that it produces (p. 37), and becomes a lever for scientific integrity.

the national infrastructure Recherche Data Gouv, which works to support the sharing and opening of research data, will continue to play a structural role in supporting researchers and providing services.

In its objectives, the SSO mentions the strengthening of open science services and infrastructures, as pillars of scientific sovereignty. Recherche Data Gouv fits naturally into the need to consolidate and structure the openness of research data, in accordance with the FAIR principles (SO.6, p. 64). This involves:

• Strengthening ecosystem competence centres
• A strengthening of the national trusted repository
• The development of the Recherche Data Gouv catalogue, indicating data deposited both in the national repository and in thematic and trusted repositories, the discoverability of which will be facilitated by AI.
   

For more information

Ministry of Higher Education, Research and Space. French Open Science Strategy. 2026.

See also the release published on the French Ministry’s website

View the catalogue : https://datatoolshub-71032f.pages.in2p3.fr/

The board currently consists of:

• Alexis Arnaud (Université Grenoble Alpes) (2022–present)
• Alicia León y Barella (University of Lille) (2025–present)
• Anthony Moalic (University of Caen Normandy) (2026–present)

"Recherche Data Gouv lends our institutions legitimacy in highlighting the importance of open research data and research data management. As a competence center, being able to rely on this ecosystem is a real asset." — Anthony Moalic

Discover the competence centers network

Antony Fornes (Research Data Valorisation Officer at the Université Côte d’Azur) and Aline Grand (Research Data Administrator at the University of Strasbourg), involved in the Communication working group of EOSC France, as well as Gilles Matthieu (Deputy Director of 'Recherche Data Gouv'), propose to dissect the main concepts relating to the EOSC to better understand the whole.

EOSC, in one (long) sentence

The European Open Science Cloud (EOSC) is an initiative that aims to strengthen the European research data ecosystem by creating a federated infrastructure for sharing and reusing FAIR data.

Head of EOSC

The governance of EOSC is tripartite. It is composed of:

a) the European Commission (Directorates-General for Research and Innovation and Communication Networks, Content and Technology) 
b) the EOSC Association, a Brussels-based non-profit organisation representing the entire EOSC stakeholder community 
c) the EOSC Steering Committee, a group of experts from the European Commission, whose members are representatives of EU Member States and countries associated to Horizon Europe.

Good to know: In France, the EOSC-France College represents the French member organisations of the EOSC Association.

The nodes : EOSC main modus operandi

The EOSC acts by mobilising specific reference communities (national, regional or thematic) or broad collaborations at European level (e.g. digital infrastructures, science hubs, etc.). These federated communities are called EOSC nodes. An EOSC node can be set up by a single organisation or by a consortium of organisations, each with its own internal governance structure and providing resources and services to the Federation. EOSC nodes must be represented within the Federation by a legal entity and comply with the decisions, rules and policies of the Federation. 

But that is not all, the EOSC is also deployed through projects. 

EOSC projects, another way to move forward together

An EOSC project is a concrete and dated object: a consortium of organisations, funded (most often by Horizon Europe, via calls from the "Research Infrastructures" cluster / INFRA-EOSC topics), with a fixed duration, a budget, contract deliverables and a coordinator. 

The EOSC Association works closely with Horizon Europe projects that contribute to the development and deployment of the EOSC; a number of European projects work in direct support of the implementation of the EOSC, while others are carried out in parallel. These projects do not build the EOSC ‘in general’: Each develops a specific brick (a service, a standard, a case of disciplinary use, a competency framework...) that then fits into the ecosystem.

Some examples of EOSC projects

• EOSC Data Commons, whose mission is to contribute to the establishment of the EOSC as the European Research Commons,
• EOSC Beyond, which aims to advance open science and innovation in research, in the context of EOSC,
• or EOSC-ARENA, centered on a sovereign generative and agentic AI environment.

Others cover digital preservation (EOSC EDEN), the creation of secure environments for sensitive data (TITAN, SIESTA), deposits in trusted repositories (FIDELIS), or the creation of a European Quality Reference Framework for Code and Research Software (EVERSE).

The concrete services offered by the EOSC

The EOSC does not function as a single portal that would host all services. It is a federation: each existing brick (national, thematic, disciplinary infrastructure) can offer services. EOSC nodes act as entry points for the various services offered within the EOSC Federation.

The basic logic is mutualization - to avoid redeveloping the same service several times, share the costs of hosting, calculation and management, and ensure that everything is interoperable (FAIR logic). The EOSC EU Node is the nucleus/model of the federation, whose role is to offer a number of common services to the entire federation. Specifically, these are the six generic services: file sharing/synchronisation, interactive notebooks, large file transfer, virtual machines, cloud containers, massive data transfer - the minimum technical base that any European researcher must be able to access. To better understand these services, see this presentation and this Couperin decryption note (in french).

Beyond this common foundation, each actor within the Federation keeps its own services, adapted to its community. EOSC brings together already existing offers, each meeting the specific needs of its field or country. Specifically, for example:

• Data Terra, selected as a thematic node for Earth and environmental sciences, brings its own tools for accessing data and software on the Earth system - built long before the existence of the EOSC, but now made accessible via the federation.
• On the humanities and social sciences side, GoTriple, resulting from the TRIPLE project coordinated by Huma-Num, has become a service supported by the OPERAS infrastructure (led by OpenEdition): a multilingual discovery tool designed specifically for SHS research needs.

And Recherche Data Gouv in the EOSC?

Recherche Data Gouv is a national service, both in terms of its scope and its missions. Its services are mainly offered to the French community. Getting involved at European level has three objectives:

1. Positioning Recherche Data Gouv at the European level. By offering services with high added value beyond our borders, the offer is more visible, gaining influence, while facilitating technical interoperability and international cooperation within a sovereign framework.
2. Facilitate access to EOSC for the community: By increasing the visibility of the services available via EOSC for researchers, but also by facilitating their support through the European network.
3. Use the EOSC lever to improve at the national level. The ecosystem then benefits from the high level of EOSC ‘standards’ to position its services at a high level of demand. In addition, it paves the way for future European funding.

Specifically, Recherche Data Gouv is currently involved in three INFRAEOSC projects:

• FIDELIS, which builds the European network of trusted warehouses (started in January 2025)
• GenAI4earth, which deploys trusted Generative Artificial Intelligence tools and models for the interdisciplinary exploitation and valorisation of Earth System scientific data within the EOSC ecosystem (started in September 2026).
• EOSC-Connect, which aims to federate about twenty national nodes of the EOSC federation (starting in September 2026)

In addition to these projects, Recherche Data Gouv is also involved in the construction of a French national node, a future candidate for integration into the EOSC federation.

Find out more

In June and July 2026, the ESR Research WG and the EOSC-France College organised a series of 3 webinars on the EOSC (in french):

• Bilan sur la stratégie politique : retour sur cinq années au sein de l'EOSC Steering Board
• Retour sur les services EOSC
• Recherche Data Gouv et EOSC : pourquoi et comment envisager la dimension européenne pour un service national

 Access to webinar replays on EOSC

Interview with the Data INRAE team

Is the space backed by a competence centre?

The Data INRAE space relies on the INRAE institutional reference centre and is linked to various local data management clusters, depending on the implementations of the submitting INRAE units.

Which disciplines are primarily represented by this data?

The most represented disciplines are agriculture, biology, food, economics and social and environmental sciences.

What is the process for submitting data? For creating a collection?

Anyone connected with the INRAE ‘identity federation’ can deposit at the root of the institutional space or in one of its three main thematic collections: Experimental-Observation-Simulation, Omics and Surveys & Texts, curation of which is provided by the Data INRAE team. The conditions for filing in other sub-collections (mainly projects or units) depend on the configuration by their administrators.

Anyone wishing to create a collection at the root of the institutional space or in one of its three main thematic collections can request it via a form accessible to INRAE agents. The administrators then manage the creation of their sub-collections and ensure their curation in accordance with the Data INRAE charter.

What curation is carried out on the platform?

In the ‘Omics’ collection, curation is done a priori. It is done a posteriori for the collections ‘Experimental - Observation - Simulation’ and ‘Surveys & Texts’. For the other sub-collections, the curation is done a priori or a posteriori (depending on their settings).

In order to harmonise practices within the space, and to increase the skills of curators, a network has been set up. Virtual classes specific to Data INRAE are regularly offered to this network. In addition, curators or administrators of Data INRAE collections can use the INRAE curation report template on the Data INRAE homepage.

Is there any integration with other repositories?

Yes, Data INRAE is linked to other Recherche Data Gouv institutional spaces or to repositories based on the Dataverse tool, in particular those of the french IRD (Institut de Recherche pour le Développement), CIRAD and CNRS, given the numerous joint research units (UMRs) and partnerships with INRAE.

Links also exist with thematic repositories, for example EMBL-EBI where scientists deposit their raw, often voluminous, data. To store their more synthetic processed data, they then prefer the Recherche Data Gouv repository, which is more general.

Do you have a model dataset or one you would recommend as an example of a ‘clean’ deposit?

A good example is the dataset of the CiTIQUE project (citizens & ticks):

Durand, Jonas; Bah, Thierno-Madiou; Lebert, Isabelle; Galon, Clémence; Carravieri, Irene; Masseglia, Sébastien; Armand, Jean-Marc; Marchand, Julien; Galley, Cyril; Chalvet-Monfray, Karine; Vayssier-Taussat, Muriel; Vourc’h, Gwenaël; Brun-Jacob, Annick; Moutailler, Sara; Bailly, Xavier; Frey-Klett, Pascale, 2026, "Dataset of the distribution of tick-borne microorganisms in human-biting ticks in France collected through a Citizen-science program (CiTIQUE)", DOI : 10.57745/FIRZ0C.

Other datasets from Data INRAE are highlighted on this page of Recherche Data Gouv portal.

Find out more

• Institutional desk: datainrae[at]inrae.fr
• URL for the institutional space: https://entrepot.recherche.data.gouv.fr/dataverse/inrae
• Open science at INRAE: https://science-ouverte.inrae.fr

• Depositing Datasets, to learn how to deposit a dataset in the repository:
  • Thursday October 8th 2026 from 2 pm to 4 pm
  • Thursday December 3rd 2026 from 10 am to 12 pm
• Administering a collection, to learn how to manage a collection in the repository:
  • Tuesday September 8th 2026 from 10 am to 12 pm
  • Thursday November 12th 2026 from 2 pm to 4 pm
• Curation, to learn how to check metadata quality of submitted datasets, in line with FAIR principles:
  • Tuesday September 22nd 2026 from 10 am to 12 pm
  • Thursday November 26th from 10 am to 12 pm
• Discovering the Recherche Data Gouv repository's APIs' , to discover how to use repository functionalities from scripts or information systems:
  • Thursday September 17th 2026 from 2 pm to 4 pm
  • Tuesday November 17th 2026 from 10 am to 12 pm
• Using the Recherche Data Gouv repository APIs', to learn the more in-depth use of APIs (Application Programming Interfaces) through practical exercises.
  • Tuesday October 13th 2026 from 10 am to 12 pm
  • Thursday December 3rd 2026 from 2 pm to 4 pm

These virtual training courses are open to everyone, with no prior registration required. A single access link can be used for all sessions:  https://webinaire.numerique.gouv.fr/meeting/signin/invite/56058/creator/29902/hash/8cd70dea21f5c3eb984d52e4214d2e35cafb616f

Throughout the year, you can also consult the Recherche Data Gouv repository guides, as well as the Recherche Data Gouv Catalogue guide (in french).

• For scientific publications: https://hal.univ-lorraine.fr/hal-05650417
• For datasets published in repositories assigning DataCite DOIs: https://hal.univ-lorraine.fr/hal-05650426

These two documents are permanently available via the footer of the Works-magnet website.

Before building these data, users are encouraged to update their institution's ROR identifier.

The submission of reliable data to the French Open Science Monitor (BSO) through these local implementations is essential for collectively improving the quality of monitoring these indicators.

The French Open Science Monitor is part of both the National Plan for Open Science (PNSO) and France's National Action Plan within the Open Government Partnership (OGP). It was developed in 2018 by the Data Engineering and Data Science team at the French Ministry of Higher Education, Research and Space (MESRE) to support and monitor the national Open Science public policy. Since 2023, its beta version has also enabled the monitoring of published datasets and software. This new feature is the result of a collaboration between the team responsible for the Monitor at the MESRE, the University of Lorraine, and INRIA.

The latest webinar recording from the BSO Users' Club is now available on the website: https://barometredelascienceouverte.esr.gouv.fr/a-propos/communication#users-club