The process of selecting where to publish data is similar to choosing the journal for publication, decision depends on your research and data type.
- Check the requirements of the funder of your research. If it is not indicated there, then choose a general data repository for your research data which could be applicable legally and ethically at your institution.
- Contact the journal publisher of your paper and consult on relevant repositories for your work.
- Does your data contain personal or sensitive information that cannot be anonymized? If yes, then consider a controlled access repository.
- If you are going to deposit a discipline-specific dataset, then consider a repository relevant to your discipline.
Here are the options to publish data:
- In a trusted repository (preferred)
Placing data in a certified repository preserves the information while providing access. Datasets in repositories are assigned stable digital identifiers and indexed by search tools, facilitating discovery and reuse. Repositories may be either generalist or subject/community specific. Nature.com recommends repositories to deposit data as part of the manuscript submission process. Use the Registry of Research Data Repositories or the Directory of Open Access Repositories to find a suitable repository, check the list of repositories on your right-hand side, or contact the library to set up a consultation.
- In Data journals or as supplemental material in a journal article
Many journals allow the publication of data or other documentation as supplements alongside the original article. This ensures that data are clearly linked to the article. There are also data journals that exist specifically to publish datasets (Data by MDPI; Scientific Data by Springer; GeoScience Data Journal by Wiley; Journal of Chemical & Engineering Data by ACS). Journal policies vary widely with regards to the type and amount of supplemental information they will publish, as well as whether supplements are subject to peer-review. Contact the journal editorial staff prior to submitting a manuscript for more information on the journal's practices regarding supplemental information.
- On a custom website
In general, publishing data to a personal website is not recommended, but there may be times when a custom solution is appropriate. For data that is of interest to a highly specific group, is only relevant for a limited period of time, or requires fine-grained access control, dedicating a specific website to distribution may be a good option.
What about providing data to colleagues by request?
Sharing data by request is a well-established scholarly practice. Data sharing is not the same as data publication, because data that is shared directly between researchers cannot be cited as easily and will not be indexed by search tools. Some repositories formally publish data but also implement access controls (ICPSR is a good example). Thus, data publication should not be confused with allowing unrestricted public access to that data. However, data publication is not appropriate in every case. Informal data exchange among colleagues is and will continue to be an important form of data sharing.