Research data objects shared with articles published in 2020 were analyzed. For each data repository cited in an article, a Globally Unique Identifier of the research data (URL or DOI) was retrieved. The identifiers were sent to the F-UJI screening tool. The screening tool uses metrics and tests to calculate a score for the completion of each FAIR principle per dataset.
The results of the screenings were aggregated at the level of FAIR principles and repositories, among others. The average FAIR score of the 299 research datasets examined is 27%. At 45% completeness, the findability of data is ensured quite well. Reusability, on the other hand, only achieves an average score of 11% due to subject-specific requirements.
The analysis of the repositories shows that general-purpose repositories – such as Zenodo or figshare – achieve a much higher FAIR score with an average of 54% than field-specific repositories – such as Gene Expression Omnibus. These repositories only achieve a FAIR score of 17% on average. However, they offer the advantage of subject-specific metadata.
The FAIR scores do not represent a conclusive assessment of the reusability of Charité research data. The monitoring of FAIR data metrics in the dashboard is a snapshot. The data management of scientists, the processing of data in repositories, but also the screening tools themselves are changing rapidly.
The interactive charts are intended to convey knowledge about the FAIR Data Principles and their monitoring and to raise awareness of the requirements of FAIR and thus sustainably reusable data.
Link to the FAIR Data Assessment on the Charité Dashboard on Responsible Research
