Guide to archiving and publishing research data

There are a number of archive solutions for research data. Here are a few tips on how to choose a suitable archive for your data.

• Explore whether there is an archive that is used in your field internationally. In some cases, the funder or publisher will require you to use a specific archive.
• It may be appropriate to choose an archive that is established in your field of study and that ensures archiving of data according to approved standards in the field. If you need help finding such a subject-specific archive, you can search the archive register re3data (re3data.org).

Norwegian archives

• As a researcher at OsloMet, you can upload your data to the OsloMet institutional archive DataverseNO (dataverse.no), which is a national, generalist archive for open research data. Guide to archiving research data in DataverseNO. Research data curators at the University Library review and curate research data before publication in DataverseNO.
• Data from surveys and data about people and society can be archived and published in the Surveybanken (sikt.no). Sikt curates all datasets that are published in the research data catalog/Surveybanken.
• For large data (single files larger than 1 TB), we recommend that you use the NIRD Research Data Archive, Sigma2 (sigma2.no). NIRD does not curate archived data and researchers must ensure their own archiving.

For DataverseNO and Surveybanken, as an OsloMet researcher, you will receive help and guidance for publishing your data there, through research support at the University Library (oslomet.no).

For Surveybanken, OsloMet also has a data processing agreement for sharing data with access restrictions, which means that you do not have to create one when archiving and publishing the project.

Other archives

• Zenodo is the EU's archive for research data (zenodo.org). Here you can archive data and other documents from all fields. Zenodo does not curate archived data and researchers must ensure their own archiving.
• Open Science Framework (OSF) is an international platform for open research (osf.io). Here you can deposit data from all research fields. OSF does not curate archived data and researchers must ensure their own archiving.

Further help with choosing an archive

Information on which criteria you should look for when choosing an archive can be found in these guides:

• How to find a trustworthy repository for your data (openaire.eu)
• Practical guide to the international alignment of research data management (scienceeurope.org)

Research data that contains personal data or for other reasons cannot be published openly can in certain cases be deposited in archives that offer access control according to defined criteria (see Sikt's research data archive, sikt.no).

Anonymized data can in some cases be deposited in an open archive. A dataset is anonymized when data cannot be traced back to an individual and the linking key has been deleted. Remember that audio recordings of voices and video material are considered personal data in themselves and must be deleted according to the information you have provided to the participants and Sikt's notification form for personal data (sikt.no).

If research data is collected with informed consent, plans for archiving after the end of the project must be included in the consent forms.

Documentation of research data is a description of what you do and have done with the data during a project. Good documentation helps to strengthen the quality of the data and your results by making the findings verifiable.

Documentation is usually an integrated part of the research method and is written while working with the data, and is summarized and attached to the data that is archived.

Documentation that is archived together with the research data can be:

• A ReadMe file: a text document that describes the content of your dataset should usually be uploaded together with the data files.
• DataverseNO: ReadMe file template (zenodo.org).
• Paradata, often in the form of a data sheet (documentation of the data creation process).
• The survey that was given to the respondents.
• Interview guide.
• Notes without personal information or identifying background variables.
• Codebook.
• Code/script that has been used, for example as a Jupyter Notebook file.

Metadata (data about data) provides additional information about your research data that makes it possible for others to find your research data, and to understand who created the data, how it was created, and where it is archived.

By creating metadata that is as complete as possible, you ensure that the data is visible and can be reused in a responsible manner. If you are working on a dataset that cannot be archived openly, you can still publish metadata.

Metadata should be readable for both machines and humans. Standardized metadata is essential for making data findable, accessible, interoperable, and reusable (FAIR). Different archives for research data use different and somewhat customized metadata standards.

The choice of archive will therefore often set guidelines for which metadata standards you should use. Overview of metadata standards (alliance.github.ui).

Good organization and a tidy file structure are important for verifiability and retrieval. We recommend that you use an archive-friendly file format and a hierarchical folder structure with descriptive and/or logical and consistent names for folders and files. You can read more about organizing research data in the DataverseNO deposit guide (uit.no).

Published datasets must have a license that determines the terms for reuse and regulates what others are allowed to do with your dataset. Many archives have their own guidelines for choosing a license, but it is recommended that you choose a standard license. Information about licenses (openscience.no).

Contact the section for research support services at the University Library (oslomet.pureservice.com), to get help with choosing an archive and support for archiving.