Feature Request - Automate Data Citations from GBIF?

[ebraker]

For every collection publishing to GBIF, GBIF tracks records that are cited via DOI:

Is there a way to ingest these as pubs/citations into Arctos? (I see Bionomia pulls this info so it seems that this could be automated)?

If so, and if the Arctos community is interested in this functionality, I think we may want to create a new citation_type value such as 'data citation' to keep these citations separate from voucher citations to indicate that no physical specimen was examined or some sort of flag that these are data-derived assertions.

At the same time, ingesting GBIF citations could create A LOT of data e.g., UCM:Herp has 490 associated publications, each of them citing anywhere from 1 to hundreds of specimens, so the data footprint might be very large.

See UCM:Herp citations here: https://www.gbif.org/resource/search?contentType=literature&gbifDatasetKey=8935e64a-f762-11e1-a439-00145eb45e9a

[dustymc]

I think this would dilute the value of citations (hard links between material or data and publication) in Arctos. The first publication in your example is a good demonstration: https://www.gbif.org/occurrence/download/0047281-240506114902167 downloaded some huge swath of data from GBIF (which includes some UCM herps for some reason) and "cited" that in a publication about insects. Any possible technical considerations aside for the moment, I can't see how mixing that in with the high-precision data in Arctos (https://arctos.database.museum/publication/10007340, for example) doesn't just muddy the waters.

[ebraker]

Yes, I noted that as well and think that some certainly are erroneous, but the second publication links 211 herp records for a paper on Thamnophis, and the majority seem to track appropriately. Perhaps there could be a tool that detects GBIF citations and then some easy thing to approve and auto-create once they are vetted? Again, that would be why I think we would also want some sort of new citation_type value to keep these citations separate from other existing citation types. I do still think there is value in the valid GBIF citations that shows how our collections support big data, biogeography and other research that would be nice to track at the record level.

[happiah-madson]

Yes, I noted that as well and think that some certainly are erroneous

I do think that sometimes its because a user downloads all of gbif and then does their filtering on their dataset. (I mean, I could be wrong, but it's my theory). We have been toying with whether we want to link these publications for exactly the reasons that @ebraker says: GBIF citations that shows how our collections support big data, biogeography and other research that would be nice to track at the record level. I'm curious to learn what other collections think.