Storing and publishing

What research data do I need to store?

At the end of a research project, the question of how to store research data arises. Ideally, it has already been determined in the data management plan which data should be retained and which should be deleted. This decision is very context-dependent and can vary from project to project. The following questions can help you decide what should be retained:

• What data are necessary to ensure reproducibility of research results?
• Does a funder or the university require that data be kept or made available for a certain period of time?
• Are the legal and intellectual property rights in place to retain and reuse data?
• Is there sufficient documentation and descriptive information to explain the data and allow the data to be found and reused?

What do I need to consider when storing research data?

Raw data

Raw data are often the basis of analyses. Based on the raw data and information about the subsequent steps, study results can be reconstructed. It is therefore important that raw data are not overwritten or changed. Raw data should therefore be stored write-protected in a separate location.

Folder structure

In the course of a research project, more and more files with different contents are created. It can be difficult to find data if there is no logical folder structure or if the file names are not clear. Therefore, how individual files and folders are labelled should be determined early on and in a uniform manner.

File naming

• File names should be constructed from various elements. Elements can be the project name, project number, research team name, type of measurement, topic, creation date, version number, etc.
• Each element should be coded so that the names are short. Approximately 25 characters is a good length for file names.
• Keeping a log file explaining the coded elements allows outsiders to recognise the files.
• Basically, files should go from generic to specific.
• Only characters from the sets A-Z, 0-9, hyphen, underscore or dots should be used in the file name. Special characters such as "% or &" should not be used, as different operating systems can assign different meanings to these characters.

Costs for storage

Depending on the size of the data sets and how long they are kept, repositories may charge a fee. These costs should be taken into account at the planning stage. The SNSF reimburses data storage costs for funded projects.

What do I have to consider when storing personal and sensitive data?

The storage of personal or sensitive data requires special technical and organisational security precautions:

• Control access to rooms and buildings
• Computer systems should be protected with strong passwords
• Personal or sensitive data should not be stored on servers or computers connected to an external network
• Access to hardware/digital copies should be logged
• Access control for data files should be implemented
• Confidential data should be encrypted before being shared with authorised persons

What are metadata?

Metadata are data that contain information about characteristics of other data:

• Information about authors, titles, keywords, or descriptions
• Information about file types, access rights, licenses, and locations
• Information about processes, actions, methods, and tools that have been used

Why do I need metadata?

Well-documented metadata have several advantages:

• Data can be found more easily
• Long-term preservation of data is promoted
• Traceability of research is increased

How do I manage metadata?

There are various standards, for managing metadata. Depending on the subject area, there are different standards.

Metadata Standards Catalog
Metadata standards by subject area  

When storing and publishing research data, the FAIR principles should be observed. Adherence to the FAIR principles ensures optimal processing of one's own research data and thus enables the data to be re-used by others for further research.

FAIR stands for Findable, Accessible, Interoperable & Re-usable.

Findable: Machine-readable metadata are needed so that research data can be found easily.

Accessible: Access to data must be regulated, authentication and authorisation must be defined.

Interoperable: In order for data to be linked to other datasets, they must be in an accessible and widely applicable format.

Re-usable: For data to be reusable, they must be comprehensible. For example, the method of data collection must be known. Furthermore, the terms of use of the data must be clearly described.

Ensure that your data are stored in a repository that complies with the FAIR data principles and that is compliant with international data and metadata standards. Examples of such repositories are re3data.org and for Switzerland we recommend DLCM long-term preservation or OLOS.