Guides to Publish Research Data and Software Following the Open Science Principles

This repository provides practical guidelines, meant for members of French research institutions 🇫🇷, to:

• publish data and software in compliance with the practices of Open Science 🔓,
• cite data and software in research publications 📖.

Links to the guidelines:

• How and Where to Publish Research Data
• How and Where to Publish Research Software

Need more help, support, advice?

• For University Côte d'Azur and CNRS members, contact DATAZUR: l’atelier de la donnée de UniCA at recherche at datazur@univ-cotedazur.fr
• For Inria members, contact the Pôle données de la recherche du Service IES at data.ies@inria.fr

---

These guides result from a presentation given in April 2026: Open Science, reproducible research, and the citation of articles, data and code alike, whose video and slides follow:

---

🔓 The Philosophy of Open Science

Open Science is the unhindered sharing of the results, methods and products of scientific research. It relies in particular on the open access to publications, source data and code.

---

❓ Why should I publish data and code?

• My research is funded by public grants, data and code have to be public too.
• It convinces reviewers that the results in my paper can be trusted.
• It allows other researchers to build on my results, compare with them (thus increasing the number of citations of my paper).
• It improves quality by forcing me to organize, document, clean my code & data.
• It enables reproducible research (see below).

Besides, considering data and code as first-class citizens in the scientific production requires that researchers should get credit for the data and code they share, and that requires tracking the citation of data and code the same way we track citation of articles.

---

🔁 Enable Reproducible Research

Too often, the results of research articles are not reproducible. One reason is the lack of details and choices (e.g. architecture, hyper-parameters, protocols), or on applicability restrictions that are not mentioned. Consequently, it is hardly possible to verify results, and therefore build upon the original work.

To enable researchers to reproduce and build on previous results:

• data and code should be published,
• on platforms ensuring long term preservation,
• under the terms of open licenses,
• assigned persistent identifiers (PID),
• and articles should refer to the data and code by means these PIDs.