Data and code

Selecting data means determining which data should be archived, published or disposed of when they are no longer needed. 

As a minimum, all the data on which a publication is based should be published along with corresponding metadata and documentation. If data cannot be published, they should be archived, and metadata should be made available online. 

The Digital Curation Centre (DCC) explains the data appraisal and selection process in more detail on its website. 

Whenever possible, data should be deposited in subject-specific repositories, as these are tailored to the needs of the field, support relevant data formats, and often provide specialized metadata. 

If you don't know a subject-specific repository in your field, search for one on re3data.org, fairsharing.org, or reach out to us. 

If there are no subject-specific repositories in your field, use a generalist one such as BORIS Portal, the institutional repository of the University of Bern, or Zenodo. 

Make sure that the repository you choose complies with funder and journal requirements. Some funders and journals also have lists of accepted repositories, e.g. 

- SNSF (non-exhaustive) 

- Open Research Europe  

- Nature 

- Plos One 

Make sure to check any restrictions or limitations of your repository of choice, for example on data types or file size. If you are in doubt, get in touch with the repository support to see if you can make arrangements. 

Do not post your datasets on blogs or institutional profile pages only, because these sites may change or terminate so that you and the public lose access to the data. By contrast, repositories help make your data findable, citable, and re-usable according to the FAIR data principles. You can link from your personal webpages to your datasets on repositories using the URLs and/or DOIs the repositories assign to your datasets. 

As part of the FAIR principles, funding bodies require a unique identifier to be assigned to the published data. When depositing your data in BORIS, a Digital Object Identifier (DOI) is assigned to each dataset. Click here for further information.

Licensing data 

The Open-Science Team recommends licensing research data under the Creative Commons Public Domain Dedication (CC0) or the newest version of the Creative Commons Attribution International Public License (CC BY) for maximal reusability. This recommendation is in line with requirements from funders such as the SNSF and the European Union. 

Licensing software 

Software developed in research projects at the University of Bern may be released as Open-Source Software if the respective research group leader approves it. Licenses used must be recognized by the Open-Source Initiative. 

Find the official guidance of the University of Bern in the directive on Open-Source Software. 

Before uploading your data to a repository, make sure you follow these steps: 

• If your raw data includes personal information, make sure the dataset you publish is fully anonymized/pseudonymized. 

• Check if your files have meaningful names so that others can understand what they contain. 

• Use open, non-proprietary file formats (e.g. csv, txt, PDF/A) wherever possible to ensure that your data stay readable over time. 

• Make sure you add sufficient documentation (ReadMe Files, codebooks, questionnaire templates, etc.) so that others can understand and re-use your data. For guidance, check our recommendations on data documentation. 

You can find detailed information and tips in our Submission Checklist. 

• It may not be possible to share some data openly due to legal or ethical reasons. This may be the case, for example, if data contains personal or sensitive personal information, or if it is subject to confidentiality agreements. However, such data can still be managed in compliance with principles such as Open Research Data and the FAIR data principles. 

• General metadata can be registered on a repository, including supplementary material containing additional information on your project and data, such as readme files, protocols, and other documentation. 

• Even if data cannot be published in open access, it may be possible to share it with individual researchers on request. In this case, no open license can be applied, but you can use a data transfer and use agreement (DTUA). 

• For information on how to do this on BORIS Portal, please check our BORIS Portal FAQ. 

When citing data, it is advisable to use either the standards applicable to the research field in question or the form suggested by the repository in which the dataset was deposited. If there are no particular standards or guidelines, the following details should be included as a minimum: 

• Author 

• Year of publication (of the dataset) 

• Title 

• Edition or version (optional) 

• Publisher (for data this is usually the archive in which the data is stored) 

• Persistent identifier (as a permanent linkable URL) 

The DOI foundation provides a citation formatter that you can use to generate a dataset citation according to a large variety of citation styles from an active DOI: https://citation.doi.org. 

Make sure that your dataset and the corresponding article(s) are linked. This helps others to locate pieces of output that belong together, and to understand them in context. 

• Add a data availability statement to your article including the dataset DOI. Some repositories allow you to reserve a DOI that you can include in the manuscript you submit to the publisher.  You can find best-practice examples for data availability statements from Open Research Europe here. 

• Add the reference to your article, including its DOI, to the metadata of your dataset. Here is an example from Zenodo (scroll down to Additional details / Related works) 

If code does not contain sensitive information, it can be shared openly via specific platforms such as GitHub. 

You can create DOIs for your GitHub releases to make it easier to reference and access them. Check the Zenodo documentation to see how to enable the integration with GitHub. 

Information and action guide for publishing open source software.