Publish data

Research data made openly available through data repositories becomes searchable, citable and gives the authors more opportunities for impact, merit and future collaborations. Publishing research data also increases transparency and reliability of presented results. It also gives others the opportunity to find and reuse existing data in new research.

When publishing data, one should follow the so-called FAIR principles. FAIR stands for Findable, Accessible, Interoperable, and Reusable, and the principles aim to make research data more accessible and easier to reuse.

Well-documented data is a prerequisite for others to be able to understand, review, and reuse them. Use established terminology and standards within your discipline. Information about data can include:

• Project level: brief information about the study and research questions, methods for data collection, how the data has been processed and analysed.
• Structural level: description of the content in the dataset - data types, file formats, relationships between files, folders and other components like code.
• Variable level: definitions and explanations of variables, values, units of measurement and other codes or abbreviations used.

The above information can also be summarised in a README file included with the dataset.

Se also:

• Suggested template for a README file (Swedish National Data Service)
• Describe, share, and preserve data (Researchdata.se)
• Best Practices for Data Submission in Generalist Repositories: A Checklist

More detailed descriptions of datasets can be published in a journal as a "data paper".

Some data should not be shared openly

There may be legal, ethical, or other reasons for not sharing or publishing research data. This may apply to data that contains personal information, trade secrets, copyrighted material, or other confidential information. In these cases, the dataset can still be described and possibly deposited in a repository. The description of the data should then include contact information and conditions for accessing the data.

If you will handle data with sensitive personal information and plan to deposit the data in repositories with controlled release procedures, you should include that information in the application for ethics approval.

See also: As open as possible, as restricted as necessary (Researchdata.se)

Data repositories and licenses

If possible choose an established subject-based data repository when publishing data. In the re3data.org register (Registry of research data repositories), you can search for data repositories in various subject areas and countries.

When publishing data, specified creators and roles may be different from those specified in publications whose analysis and results are based on this data. For examples of roles see CRediT – Contributor Roles Taxonomy.

Examples of subject specific repositories

• CESSDA – social science.
• DARIAH – humanities.
• ELIXIR deposition databases for biomolecular data
• FEGA Sweden - a repository with controlled disclosure procedures for sharing genomic data.
• HEPData – high-energy physics.
• NOMAD – material science.
• PANGAEA - geosciences and biological sciences.
• Swedish Biodiversity Data Infrastructure – biodiversity.
• SciLifeLab Data Repository – Life Sciences.
  

Examples of interdisciplinary repositories

• Swedish National Data Service (SND) - a certified repository
  Data sets published in the SND Repository are reviewed by staff at the Uppsala University Research Data Support in dialogue with responsible researchers. The aim is to make published data more FAIR.
  
• Zenodo - a repository operated by OpenAIRE and CERN, funded by the European Commission. Zenodo can be used for storing and publishing data sets, code and publications. Code created in Github can be published via Zenodo and be assigned a DOI-number.
  
• Figshare - a data repository run by the company Digital Science.
  
• Dryad - repository run by a network of universities, scholarly societies and publishers.
• Dataverse - a data repository at Harvard University.

Licenses for data and code

To facilitate the reuse of published data, one can choose a license and specify conditions for reuse, for example, using Creative Commons licenses or Open Data Commons. The chosen license can inform about wethether modifications and commercial use are allowed, and how the data can be shared further.

For data developed within research a license that allows recognition to those who created the dataset is preferable, see more about licensing research data at Researchdata.se

Some links about licenses for code, computer programs, and databases:

• To choose a license for open source software, GitHub has created the tool Choose a license.
• Licensing Assistant – a tool to find and compare software licenses.
• Open Source Initiative lists licenses for software and has a FAQ on the subject.
• European Union's open source license
• Open Data Commons Open Database License (ODbL) – An open license for databases, from the Open Knowledge Foundation.