search

gbif / literature-ws

Apache License 2.0
0 stars 1 forks source link

Enable search by DatasetID #10

Open MortenHofft opened 1 year ago

MortenHofft commented 1 year ago

This issue is created based on a feature request from @dagendresen - Dag I hope I relay your feature request properly?

We have recently added datasetName and datasetID to occurrence search as a way for publishers to organise their datasets into smaller chunks or associate records across datasets.

That prompts the next request, how to count citations for those datasetIDs? So adding something similar to https://github.com/gbif/literature-ws/issues/8 and https://github.com/gbif/literature-ws/issues/9

But where those are relatively simple in that they rely on the downloads having information about the datasetKeys, then this is different and presumably require us to store all seen datasetIDs against each download. Which if misused in the data could be millions (imagine someone added a new datasetID per record, just like we see for eventID).

At least it seems very different to me, but I might be wrong.

dagendresen commented 1 year ago

I believe that data citations in GBIF today are tracked only on the dataset level. My question in the hosted portal meeting was (1) a request to generate data citation tracking to a network (such as OBIS and Living Norway) in a similar manner as is made today for data publisher institutions. And second (2) if it is possible to generate data citation tracking to a project through the project ID in the EML metadata on the dataset level. today many metrics are tracked on the project level but only for the GBIF BID, BIFA and CESP projects described through this project ID in the EML - but I believe not for other projects we describe in a similar manner from the nodes?

The other alternative, as was described by others in the meeting, to explore if the datasetID on record level can be used for tracking data use citations is indeed very interesting!! Overall also to pull data citation down not only to record-level -- but ultimately to object type level such as e.g. citation for individual collection specimens or individual monitored organisms inside of the simple Darwin Core denormalized records is a very noble goal, and would be extremely interesting if solved! Maybe (?) the new unified data model can help here? I do not think that such object-level citation tracking goals can be solved with the simple Darwin Core Occurrence model in use today?

dagendresen commented 1 year ago

I believe that adding projects to the registry to supplement networks could make the very same type of possibilities as the GRSciColl links for collections!

(Here not thinking of datasetID on record-level but only on the projectID in the EML dataset level metadata -- but maybe also an ID-name in datasetID matching a pre-registered projectID in a registry list of projects nodes want to track could be tracked as well...).

MortenHofft commented 1 year ago

Thanks Dag

today many metrics are tracked on the project level but only for the GBIF BID, BIFA and CESP projects described through this project ID in the EML

We do not track literature for projectID, not for BID, BIFA etc either. But I believe there is a feature request somewhere for it