When studies published in scientific journals make use of GBIF-mediated data, we log how they cite GBIF–i.e. either 1) using the recommended format that includes a DOI, or 2) using a more generic format (e.g. simply mentioning the GBIF.org website).
So how are publishers doing? The following table shows citation type broken down by journal publisher:
|Publisher||DOI||Generic||Grand Total||% DOI|
|Public Library of Science (PLoS)||10||43||53||18.87%|
|Oxford University Press (OUP)||5||33||38||13.16%|
|The Royal Society||2||11||13||15.38%|
|Cambridge University Press (CUP)||4||7||11||36.36%|
|Proceedings of the National Academy of Sciences||1||6||7||14.29%|
|Informa UK Limited||2||5||7||28.57%|
|Frontiers Media SA||2||5||7||28.57%|
|Walter de Gruyter GmbH||2||4||6||33.33%|
|American Association for the Advancement of Science (AAAS)||1||5||6||16.67%|
A few caveats:
- we’ve only been logged this data consistently since August 2017 so this data represents only about 25 per cent of all papers using GBIF-mediated data
- only publishers with more than 5 known papers are included
- studies that access data directly from the API and via an API-based script (such as rgbif, dismo, spocc, etc.) obtain data without triggering the generation of a DOI. As such, they don’t have DOI to cite but this distinction isn’t made in the data shown. These are included in the “generic” count.