New Kids on the Block
/Ted Habermann, Metadata Game Changers
PIDFest 2026 brought together many old friends in Leiden with conversations covering a wide variety of old and new topics. While it was great to see the old guard, it was also wonderful to see new faces, many attending their first or second PID focused meeting. These new faces made me wonder about how new DataCite members are using DataCite.
To provide context for understanding the new repositories, I collected completeness data for 3,807 active DataCite repositories. These data are available as a new version of the Metadata Completeness for Four FAIR Use Cases for all DataCite Repositories dataset.
Big Picture
The big picture includes measurements of metadata completeness during January (orange in Figure 1) and October 2026 (blue in Figure 1). The two measurements have no significant differences in the medians. Figure 1 also shows the new repositories (green) that are discussed below.
Figure 1. Distributions of FAIR completeness by use case for three groups: all repositories 2026-01 (orange), all repositories 2026-10 (blue), and all repositories measured for the first time during 2026 (up to Oct. 20). Medians are marked as horizontal lines and labeled.
New Repositories
Violin diagrams are good for comparisons across distributions, but looking at all the underlying data can also provide helpful context. Figure 2 shows the total FAIRness for 413 repositories with measured FAIRness for the first time in October 2026, the violin on the far right in Figure 1. Most of these repositories are still small, with 378 (92%) having less than 1000 records and a median size (20) much smaller than the overall DataCite median (135). The largest in this group is the List of Prokaryotic Names with Standing in Nomenclature (LPSN) with 132,525 records (5000 sampled) which migrated to the German Collection of Microorganisms and Cell Cultures GmbH at the Leibniz Institute during 2026. It is shown on the right side of Figure 2 along with other repositories with > 5000 records.
Figure 2. Total FAIRness for repositories measured for the first time during 2026. Several outstanding examples are identified by repository id and listed in Table 1. Repositories with more than 5000 records were sampled to 5000.
Forty-one of the new repositories have total FAIRness above 40%, the informal “bright spot” level proposed by Habermann and Robinson (2026). Many of these are clustered at the left side of the plot with < 5 records (Table 1). The LPSN mentioned above (zypi.gzsawm) is a large repository bright spot with FAIRness = 42.6%. All except two of the repositories with between 1000 and 5000 records (or samples) show completeness below the median; hbvj.abts and nasasomd.nlsp have FAIRness above 33% (Figure 2).
| Repository Name | Repository ID | #Records | Total FAIRness |
|---|---|---|---|
| SINTETIC catalog | rpak.kshvcr | 2 | 52.5% |
| REPOSAR | rpak.pjpaab | 2 | 52.5% |
| Institute of Archaeology of the Czech Academy of Sciences, Prague - spatial data | arch.sdsvvf | 1 | 50.8% |
| NAKI III. Národní a kulturní identita českých komunit na Ukrajině (National and Cultural Identity of Czech Communities in Ukraine) | nacr.naki | 1 | 50.8% |
| FLEXPART | qpcj.flexpart | 3 | 50.3% |
| Leibniz Institute DSMZ – German Collection of Microorganisms and Cell Cultures (LPSN) | zypi.gzsawm | 132,525 (5000 sampled) | 42.6% |
| Aberystwyth Research Portal | hbvj.abts | 1319 | 34.2% |
| NASA Life Sciences Portal | nasasomd.nlsp | 2675 | 32.7% |
Overall, the median FAIRness of the new repositories is 26%, higher than the overall median observed during January and October 2026. Figure 1 shows the differences observed for all four use cases. Several use cases showed interesting increases in completeness.
The largest difference in completeness for the new repositories was in the identifier use case which includes ORCIDs for creators, RORs for creator affiliations, funders, and publishers, and award numbers for funding awards. Figure 3 shows the distribution of Identifier completeness for all 413 new repositories.
This use case includes 18 metadata elements and the generally small numbers of records (Figure 2) results in a distribution that has clear peaks associated with integer numbers of elements included in the metadata, i.e. 1/18, 2/18, … 15/18. The plots across the top of Figure 3 show identifier completeness snapshots of 14 repositories, each illustrating steps along the many possible paths from 6% to 83% identifier completeness. The first repository on the left includes only one of eighteen elements, the ResourceType, in all records, i.e. 100% completeness. As we move up the identifier ladder, various elements are added one by one to reach 15 in the most complete repository on the right. The plots also reflect the fact that several identifiers include three metadata elements: the identifier itself, the identifier type, and the identifier schema, see list of metadata elements on the right side of Figure 3, so they tend to occur in pie slices of three related elements.
The other three use cases in the complete suite (Text, Connections, and Contacts) also offer multiple pathways to metadata completeness that can be selected depending on repository priorities. Completeness details across all use cases can be explored for this repository, or any of the bright spot repositories, using the links in the ID column of Table 1.
Figure 3. Histogram of Identifier Completeness for 413 repositories that were measured for the first time during 2026. This use case includes 18 metadata elements listed on the right. The histogram has peaks for completeness values that correspond to integer numbers of elements (1-15). The radar plots show examples of repositories in the group that correspond to these values to illustrate how the plots reflect increasing completeness. Several of the bright spots listed in Table 1 (nasasomd.nlsp, zypi.gzsawm, and nacr.naki) are included.
Conclusion
Repositories measured for the first time during 2026 are generally small in size, but, on average, they have more complete metadata than the overall DataCite average. Increases in completeness > 1 percentage point were observed in two of the four FAIR use cases, with the largest difference in the identifier use case which increased by 7 percentage points (from 19% to 26%). This bodes well for identifier use across all of DataCite. Organizations creating the new repositories are following existing good examples and using identifiers to unambiguously identify authors, generally with ORCIDs, and several kinds of organizations with RORs in their metadata. Forty-one (10%) of the new repositories have completeness >= 40%, the bright spot level originally defined by Habermann and Robinson, 2026.
At the same time, all but two of 20 repositories with between 1000 and 5000 records have completeness which is less than the overall median. The community needs to find more effective approaches to help new members understand and take advantage of the broad capabilities supported by the DataCite metadata schema.
References
Habermann, T. and Robinson E. (2026). DataCite Bright Spots – Repositories, Consortia, and Improvements. Front Matter. https://doi.org/10.59350/v2may-69s52