Criteria for releasing a data publication

This information is meant as a guideline for the decentralized data managers (DDM). But it can be also helpful for the authors to ease the work of their DDM. Before pressing the button publish or review, check the following points:

Formal criteria

Check the metadata fields.

  • Several metadata fields have text boxes like authors or keywords. Click on the plus or minus to add or remove text boxes.
  • Check, if there is only one term filled in per box.
../_images/box_author.png
  • The names of authors should always be filled in identically, e.g. last name, first name.
  • Is every author assigned with his/her affiliation?
  • Whenever existing, identifiers like ORCIDs should also be indicated.
  • Is the FZJ metadata schema correct?
  • Check the spelling.
  • Randomly the content can be checked, e.g. is there a content at all in the file and does it look like scientific data.
  • If appropriate, use open file formats: Recommended Formats Statement

Content-related criteria

How to improve the quality of metadata?

To improve the quality and reusability of the dataset, it is recommended to fill in the optional metadata fields.

  • Use the subject-specific metadata schemas which are provided in Jülich DATA: Metadata References, Engmeta and EnzymeML.
  • Keywords help to make the dataset findable. Use a controlled vocabulary (intranet) , if there is a standard in your community.
  • Give credit to all the people who were involved. Therefore are the fields producer, contributor and distributor.
../_images/credit.png
  • The CRediT project developed the Contributor Role Taxonomy (CRediT) which defines types of contribution in science.
  • Indicate provenance data like production place, production date, grant information, etc.

External data sources

  • If external data sources are used, check their licenses and take care to grant your dataset with a compatible license. If necessary reject unlicensed data sources.
  • Refer to the external data sources in the metadata.
  • Use the metadata fields “Related Datasets” or “Related Material”. You find it under the Citation Metadata schema. If there is no DOI available, it is possible to fill in an URL/URN instead.
  • Indicate the external data sources in the “description” field, especially if they under a license which requires to credit the author by editing.

README File

As it is common in the open source community to provide a README file, the same can be done for a dataset. Many guides and templates can be found in the internet. The one linked here is from Cornell University:

Guide to writing “readme” style metadata

Quality indicator for research data and research software publications

The Task Group Helmholtz Quality Indicators for Data and Software Publications developed a quality indicator for research data and software publications.

Find more information in the paper Towards a Quality Indicator for Research Data publications and Research Software publications – A vision from the Helmholtz Association.