Track dataset DOIs and their references

[MathewBiddle]

What should we add?

@7yl4r started some great work trying to track DOI citation counts. To get a sense of how IOOS data are used and referenced, it would be interesting to see if this process works for dataset DOIs.

Through OBIS (https://obis.org/institute/23070) we can get dataset DOI's from GBIF in an automated way. That would probably be a good start for now.

Other dataset DOIs would probably be a manually curated list from NCEI as I don't think the RAs are minting dataset DOIs.

Reference

• doi citations counts marinebon/data_reports#2

[MathewBiddle]

from ioos_metrics import ioos_metrics

#| code-summary: fetch the latest mbon data
stats_df = ioos_metrics.mbon_stats()

stats_df['doi'].unique().tolist()

['10.15468/fxjpbr',
 '10.15468/rvlbkl',
 '10.15468/sqpu7z',
 '10.15468/hczoxz',
 '10.15468/nvvifc',
 '10.15468/yg2y7v',
 '10.15468/mrxge2',
 '10.15468/3ifpo0',
 '10.6073/pasta/d302929b97723a1425364e1a19efbf55',
 '10.15468/buqg4u',
 '10.15468/ykbf7p',
 '10.15468/yvgzjt',
 '10.15468/rdkfyf',
 '10.15468/gaekez',
 '10.15468/nuqkih',
 '10.15468/uzpt9m',
 '10.15468/g8q8ey',
 '10.15468/dwxlan',
 '10.15468/quejlc',
 '10.15468/419say',
 '10.15468/zq1ep2',
 '10.15468/es1iso',
 '10.15468/kfnaep',
 '10.15468/1pyhh5',
 '10.15468/dple14',
 '10.15468/pcikkj',
 '10.15468/rexjmu',
 '10.15468/jlkkrw',
 '10.15468/t0r3vt',
 '10.15468/dfyb57',
 '10.15468/ranfdh',
 '10.15468/6chrsz',
 '10.15468/7dnpl0',
 '10.15468/84ntea',
 '10.15468/oomxex',
 '10.15468/7zofww',
 '10.15468/vnvtmr',
 '10.15468/06aqle',
 '10.15468/adis7b',
 '10.15468/tnn5ra',
 '10.25494/p6js3m']

[ocefpaf]

@7yl4r started some great work trying to track DOI citation counts.

Tylar, what are you using to do this? Any existing library or something you create?

[7yl4r]

@7yl4r started some great work trying to track DOI citation counts.

Tylar, what are you using to do this? Any existing library or something you create?

OpenCitations is doing all the fancy stuff [stack overflow overview].

All I've done is parse the json response [R code here].

[ocefpaf]

I'm not able to create an access token. Look like they believe I'm a bot :-/
Are you using one @7yl4r ? Did it work for you?

[ocefpaf]

Got one, had to change my ISP. A quick test doesn't show interesting results though, no DOI in that list worked for me:

https://gist.github.com/ocefpaf/b46f2edf8230cb698652ef9586616d4d

[MathewBiddle]

Interesting. When I go to one of the dataset DOI landing pages (eg. https://dx.doi.org/10.15468/fxjpbr) you can see there are 42 citations (https://www.gbif.org/resource/search?contentType=literature&gbifDatasetKey=f4b56e69-4ff5-4cd2-bc48-ec273232d9e0). You can download them as a tsv here https://api.gbif.org/v1/literature/export?format=TSV&gbifDatasetKey=f4b56e69-4ff5-4cd2-bc48-ec273232d9e0

As I look at the papers citing that dataset, I can see most of them are using the entire GBIF download, not the individual dataset. For example, this paper cites GBIF. 2022. Global Biodiversity Information occurrence. Download from 1 March 2022. doi: [10.15468/dl.nqbg5v](https://doi.org/10.15468/dl.nqbg5v).

So, this makes me think that the citations list at GBIF is manually curated. Interesting. I've learned something and, as expected, things are more difficult than anticipated.

@sformel-usgs do you have any insight into how the number of citations is computed at GBIF? It looks like it might be manually curated?

[ocefpaf]

Just for fun I tried Google Scholar via the scholarly library:

https://gist.github.com/ocefpaf/f6a65876874f6ee3065a9f18dda49719

Only two DOIs are found and both have lower citation numbers than their actual page.