How Many When (All Time)

Cite this blog as Habermann, T. (2022). How Many When (All Time). Front Matter. https://doi.org/10.59350/zmapp-xsb40

I recently introduced a tool designed to help measure and understand DataCite metadata usage and showed how it could be used to answer a question about usage of new DataCite Resource Types.

I extended that question to all resource types in the last blog, but the answer was incomplete because DataCite facets are generally limited to the top ten items (see question from Matt Mayernik and comment on blog). In order to overcome that limitation, I added a new capability to the tool to do retrievals for all of the years that DataCite has been registering resources, 2004–2022. The command for generating these data is:

retrieveDataCiteFacets --years --htmlout -fl resourceTypes --facetdata

Running this command with the –facetdata creates a file named DataCite_years_resourceTypes__timestamp.csv which gives the complete history of all twenty-eight resource types for nineteen years. I was surprised to see that the resourceType facet is not limited to ten items – which makes it possible to see the history of all resourceTypes at once. A great exception to the rule.

Figure 1 shows the count of records and resourceTypes through the entire DataCite history. This number has been rising steadily through time with the largest increase (1.7 million records) last year. It looks like that increase will be dwarfed this year with 4.3 million new records registered during the first five months of 2022.

Counts of DataCite resourceTypes vs. Time.

Figure 2 shows the yearly distribution of objects by resourceType. I noted earlier that so far this year, there were more Text resources than Datasets. The data shows that this is not that unusual with Text resources outnumbering Datasets during five of the eighteen years.

Many of these resourceTypes show rather heterogeneous histories. For example, note large influxes of Images during 2010 and 2019, PhysicalObjects during 2017, and Preprints during 2022. These provide interesting stories for further investigation and understanding.

Distribution of DataCite resourceTypes / year. The five most common types are labeled (P. Obj. = PhysicalObject, Pprint = Preprint). Legend the same as Figure 1.

All of the features of these plots reflect complex decisions made by DataCite members as representatives of the broader research community. Some other interesting stories of Data history are The Tenth Anniversary of Assigning DOI Names to Scientific Data and a Five Year History of DataCite and 10 years of DataCite – How it all began.