FAQ on the guideline for handling research data at LUH

Research data are data generated and processed in the course of an academic working process. They form the basis of research results and can take many different forms, such as measurement results, secondary analyses, visualizations, models, or the results of surveys and data collections. The possible file formats for storing numbers, text, program code, or graphics are just as diverse. Sometimes even physical samples such as minerals or tissue are defined as “data.”

The type of data and their file formats determine whether and how they can be processed automatically, how compatible they are with other hardware and software systems (Interoperability), and whether they remain readable in the long term (Re-usable). File formats should therefore be as open and widely used as possible.  

However, some devices and many applications save data by default in a manufacturer-specific format that can only be read with special software. In such cases, check whether you can export or convert the data to an open format to improve interoperability.

Recommendations on file formats can be found, among other places, on the website of the State Initiative for Long-Term Availability of the State of North Rhine-Westphalia.

Folder and file names should consist of elements that allow the contents to be quickly identified. For example, they can provide information about the creation date, the file version and the person who edited it. These elements should follow a standardised format and be strung together in a consistent order. Ensure that naming conventions are agreed in advance, set out in writing and adhered to throughout the research process.

The more information file names contain, the longer they can become. As some programmes cannot process very long file names, it is advisable to include the parts of the information that are the same for all files in a folder in the folder name instead.

Tips for naming files

• Date formats in the format YYMMDD, e.g. 150828 for 28 August 2015
• Abbreviate personal details, e.g. to initials
• Only use the following characters for file names: A-Z a-z 0-9 _ (underscore)
• Do not use umlauts, spaces or special characters, as many programmes do not recognise these characters as part of the file name or do not display them correctly

When you collect, gather and process data for your research, you should always document where they come from or how they were generated, as well as the processing steps they undergo. Structured information used to document and describe research data is known as metadata. This includes, for example, details of the research context and technical characteristics. Metadata can be collected in various ways during the course of a project, e.g.:

• in a database
• in an electronic laboratory notebook
• in a Git repository (version control)
• in spreadsheets (e.g. Excel)
• in Readme files (text, PDF)
• in a structured XML or JSON file
• embedded within the data files (e.g. in the file header)

Good, structured documentation, including the recording of metadata, makes the data...

• easier to find (including by machines, e.g. via search engines)
• easier to understand (or understandable in the first place) and thus reusable
• citable and thus attributable to the creator of the data
• more valuable to science, as content, quality and processing status can be better assessed

Wherever possible, this basic descriptive metadata should always be included when data is published:

• Unique identifier (e.g. a DOI)
• Title of the dataset
• Creator (principal investigator)
• Collection date (including versions)
• Format (software required, if applicable)
• Subject area
• Description of the data / Abstract
• Data collection (spatial / temporal)
• Organisation
• Rights / Licence terms
• Relationship to other objects (data, texts, etc.)

Metadata schemas define the exact format of this information to facilitate machine processing. Particularly widespread general (generic) standards include DataCite and Dublin Core. Further information on this topic can be found, for example, on the website of Cambridge University.

Plans for managing research data should be set out in a data management plan (DMP) at the start of the research process. A DMP primarily includes the following:

• Administrative details of the research project
• What data will be used in my project? (Data generated by the researcher, pre-existing data)
• How and where will the data be stored? (File names, storage locations, backups)
• How and using what methods are the data and their processing documented? (Documentation systems, metadata standards)
• How are the data processed? (Tools and methods)
• What legal and ethical aspects must be taken into account? (Data protection, rights of use)
• How is data made accessible? (Archives, repositories, PID, licences, access restrictions where applicable)
• Who is responsible for which data management tasks within the project?
• What resources are available for data management? (Funding, infrastructure, staff, time)

In general, you should draw up a data management plan for internal project use in as much detail as possible. Adapt it accordingly if the research process deviates from the original plan or if certain aspects need to be clarified. You can find our detailed recommendations on the structure and content of DMPs here. On the ‘Tools’ subpage, you will also find online services that can assist you in drawing up a DMP.

Further information

• LUH Guide to Drawing Up Data Management Plans (German)
• How to Develop a Data Management and Sharing Plan, Sarah Jones (DCC)
• DCC Checklist for Data Management Plans
• DFG Checklist for Handling Research Data

Before you collect, process and publish scientific data, you should check the legal framework and guidelines governing the handling of research data. Personal data, for example, are subject to data protection laws. If you are processing data collected by other individuals or institutions, check in advance whether, and under what conditions, re-use is permitted. Has the data perhaps been published under specific licences? You should also clarify with your employer what rights you have to use the data you have collected. If you are generating data jointly with others, it is best to draw up written agreements in advance, setting out who is permitted to use which data and in what form. Detailed information, points of contact and further links on legal and ethical issues can be found here.

The loss of your data – which has been collected at considerable cost, time and effort – and the analyses based on them could have significant negative consequences for your research. An automated backup (ideally on a daily basis) to storage media located in at least a different building than the original is therefore essential.

In an ongoing project, you should use departmental servers automatically backed up daily via the LUIS Backup & Restore service for storage, if possible. Alternatively, LUIS provides a ‘Projekt-Seafile’ cloud storage as part of the ‘Projektablage’ bundle, which is also automatically backed up daily.

If you are working with sensitive data, you should restrict access, as well as read and write permissions, to the immediate circle of staff members. Such restrictions can be set, for example, via group policies at file system level or through the relevant settings in file services such as the Projekt-Seafile provided by LUIS. Free commercial cloud storage services and unencrypted USB drives are not suitable locations for sensitive data.

For data protection reasons, you must always store personal research data in encrypted form. You can encrypt entire file systems on mass storage devices such as hard drives and portable USB drives to prevent unauthorised access (disk encryption). Most operating systems, such as macOS, Windows and Linux, already come with software for this purpose (FileFault, BitLocker, dm-crypt). A good alternative is the open-source programme VeraCrypt. The software allows encrypting not only entire drives and partitions but also directly of individual folders. The latter also works with a number of archive managers (e.g. 7-Zip). With Cryptomator, you can create encrypted file containers, which can then be securely stored in cloud storage services such as the Projekt-Seafile. LUIS provides further information on its page on data encryption.

In line with good research practice, research data underpinning published findings must be retained for at least ten years for verification purposes. One way to do this is by storing the data in a professional data archive. For LUH members, LUIS provides such an archive.

Where no legal or ethical restrictions apply, data should not only be retained but also made as openly accessible as possible. Researchers can meet this expectation by uploading their data to a research data repository or data centre. Once published there, third parties can re-use and cite the data. Most repositories have specific requirements, which researchers should ideally consider before creating their data. These requirements typically include some or all of the following:

• Use of open data formats to facilitate long-term archiving and access
• Provision of mandatory metadata to improve discoverability and usability
• Assurance that archiving and access comply with copyright and data protection laws
• Assignment of a licence with clear terms for third-party reuse

LUIS operates an institutional data repository. However, subject-specific repositories should be preferred, if available. You can find an overview of existing data repositories at www.re3data.org.

What to consider when choosing a repository

• Guaranteed data retention for at least 10 years
• Metadata registration for each dataset, meeting at least DataCite or Dublin Core standards
• Assignment of a unique, persistent identifier (e.g., a DOI) for every uploaded dataset

Many funders, universities and scientific organisations require or recommend that research data and other results are openly accessible after a project ends. This makes it easier to evaluate research findings and to enable other researchers to reuse data.

However, it is neither possible nor useful to publish all data generated during the research process. Data worthy to be published may be all data that are needed to understand a project’s outcome. The most important criteria are as followed:

• Uniqueness: No duplicates of the data have already been published elsewhere.
• Extremely limited reproducibility: The data cannot be re-generated or only at great expense.
• High professional relevance: The data is of particular interest to your professional community or even across disciplines.
• Basis of text publications: You have published a book or article based on the analysis of this data.

To ensure that your data is reusable, please note the following:

• Adequate documentation: Provide sufficient descriptive metadata so that the data set can be searched in a database (e.g. a repository).
• Readability: If possible, save the data in open, widely used formats that can be opened independently of platforms and does not require special (possibly not permanently available) hardware and software.
• Rights: Check whether the rights of third parties may prevent publication (e.g. copyrights or personal rights). If this is the case, try to have all necessary rights granted in writing by the persons concerned. Provide your data with an open license (e.g. CC0) so that it can be used by anyone without restrictions.

By assigning a licence, authors define the terms under which third parties may reuse their data. We recommend using the most open licences possible for data publications, as these impose few or no restrictions on reuse. This allows others to use, modify, and redistribute the data as they see fit. Some licences do restrict these rights, but these are no longer considered "open". Most repositories require choosing a standardised licence before accepting data for publication.

Recommended licences

• Open Data Commons licences
  Licences for open data and databases
• Creative Commons licences
  Suitable open licences for research data include:
  • CC0 (no reuse conditions)
  • CC-BY (attribution required)
• Data Licence Germany (Datenlizenz Deutschland)
  A licence specifically designed for the German legal framework. It aims to standardise and simplify licensing for users in Germany.
• Software licences
  A software licence determines how others may reuse source code.
  • Overview of various open-source software licences
  • Guide to choosing open-source licences