We’re excited to announce a new data citation API endpoint and are seeking your feedback. The new service makes existing data citation relationships in our metadata available, thereby surfacing this part of the research nexus. At the same time, we’ve decided that it’s time to move on from Event Data.
An archive of Event Data events collected while the service was running is now available on Academic Torrents, alongside our public data files. See https://doi.org/10.13003/wjyr-rv9j.
The total archive size is 107.6 GB (when compressed) and includes all events generated by agents run by Crossref, a total of 1,827,665,106 events.
Feedback on Data Citations APIs
Dear Crossref-Team,
thank you for your efforts to improve the monitoring of data citations! As a research data steward within a research institution, I am very interested to know how many of your published datasets (which we register via DataCite DOIs) are being reused, i.e. cited by other researchers.
I would like to give feedback on the Data Citation API’s beta version. Sorry, if this is the wrong thread!
Approximately a year ago, I’ve checked the citation counts of selected data publications via the Events Data which worked well. Today, I wanted to repeat this but realised that the API has been exchanged by your new service. So, I tested the beta-version of the new API with the following DOI: 10.3220/DATA20200203151139. In the past, i.e. using the Event Data function, I found multiple resources which cited this dataset. However, the beta version now renders “total-results”: 0.
Could you help me? Am I expecting something that the beta version should not deliver? Or does it have to do with the DOIs of the objects which cite the chosen dataset?
Thanks for your interest in data citations! The DOI that you mentioned is registered with the mEDRA DOI registration agency. The scope for the data citations API currently only includes DOIs registered with Crossref and DataCite, because we can more easily check the type. I’ve had a look at this DOI in the mEDRA API and there’s nothing to indicate that it’s a dataset, which means that we can’t know that a reference to this DOI is a data citation.
The scope of Event Data was broader and included references to and from any DOI. This made the dataset much larger, but also caused us problems with stability and reliability. The scope of the data citations API is deliberately narrower to satisfy more specific use cases and to make maintenance of the service more manageable.
Thank you so much for the elaborate and helpful answer! I have tried again, now with a subset of publications whose DOIs were registered with DataCite. This rendered datasets which were cited at least once, so I got the expected results.
Hi Mrittman, thank you for the link, but the link seems not be working. Is there any other way that we can have access to the historical event data? and I would like to express my sincere admiration to your work on crossref event data. It is pity to see it gone.
Thanks for your interest. The link seems to be working for me - https://doi.org/10.13003/wjyr-rv9j. You could try the direct link at Crossref Event Data Archive - Academic Torrents if that doesn’t work.