GRSciColl - explore how to link GRSciColl to Bionomia

[ManonGros]

This is a topic that came up during the Technical Support hour for Nodes.
In order to show case collections by specific collectors, it would be good to explore if we could have a good way to look GRSciColl to Bionomia.

It could be through the important collector field or the GBIF specimen-related occurrences annotated in Bionomia. Or even through Wikidata? #471

[dshorthouse]

There are a number of directions this could take, but there's risk of feature creep depending where new enhancements are made and what might be their purpose.

Do you suppose the request here might have been to streamline communications between volunpeer "Scribes" in Bionomia and the responsible parties behind the individual occurrence records? We often stumble across inconsistencies in data & while actively curating links in the Bionomia context, it'd be nice to flag an issue right there and then. While not immediately relevant to GRSiColl, a new term in DwC would be very handy for this purpose if it existed: tdwg/dwc#180. Lifting this into a formal proposition has stalled.

If on the other hand there's a desire to include some notable collectors in GRSciColl, this sounds like a job for a wikidata query, which would have the added benefit of that community actively curating registry data. For this to work, we'd need GRSciColl IDs in wikidata (or some combination of shared reuse of RoR, GRID), use of the "collection items at" property, plus the Bionomia ID property, all used in concert. The outcome here would be a repeatable SPARQL query that would allow a cross-walk between an entry in GRSCiColl to pull a handful of collectors for display.

Happy to discuss these or others, whatever might be the direction.

[scooleman]

Creating a resolvable link from a filled-in value for 'important collector' in GRSciColl to that naturalist's Bionomia profile seems like a very logical technical solution at the moment. However, it implies a few remarks since Bionomia:

• fetches GBIF data records based on just two data fields, namely collector and identifier. That might mostly cover the majority of the naturalists' specimens, but could lack a notable part, if already identified specimens were acquired by donation, exchange or purchase. Naturalists could thus also be just donator.

• aggregates all linked GBIF data records from different institutes. That seems to be very 'broad', if an institute wants to highlight a famous collection from its own Scientific Heritage.

The Dautzenberg Mollusc collection at RBINS, the Institute of Natural Sciences in Belgium, is for instance famous as part of one of the three largest shell collections in the world. Philippe Dautzenberg is therefore mentioned in the data field 'important collector' at RBINS GRSciColl collection of Molluscs (to where RBINS Mollusc collection data on GBIF will be mapped if the collectionCode is appropriately updated and republished (in prep.). Currently, the Dautzenberg data records are not easily retrievable in GBIF web interface (except for relatively advanced data users) due to several values in the fields:

• collection code: several representative (but soon to be modified) collection codes (per dataset);
• collector OR identifier:
  . with spelling variations of the naturalists’ name;
  . lacking records of specimens that the naturalist only donated.

Dautzenberg's Bionomia profile might be great, but we're still searching to make all the RBINS Dautzenberg data records easily citable and retrievable, preferably in GBIF web interface, for all data users.

[dshorthouse]

Dautzenberg's Bionomia profile might be great, but we're still searching to make all the RBINS Dautzenberg data records easily citable and retrievable, preferably in GBIF web interface, for all data users.

Limitations in role or action notwithstanding, this is the ultimate purpose of dwc:recordedByID. The challenge that Bionomia is trying to solve is how to populate this term from the sources where Dautzenberg's specimen records are shared, regardless of how his name was spelled. For what it's worth, these are now shunted to Zenodo as a versioned archive, https://doi.org/10.5281/zenodo.8030829 - perhaps useful for the present use-case.