18617
Reporter: kylecopas
Assignee: jlegind
Type: Bug
Summary: No country listed for ICIMOD --> no data from NP
Status: Open
Created: 2016-06-29 11:29:09.207
Updated: 2016-06-29 14:57:01.129
Description: There's no country listed in the registry for ICIMOD, which means that their 14 datasets aren't counted as coming from Nepal in the country report (see http://www.gbif.org/country/NP/publishing et al.).
Can this be adjusted prior to the imminent update of the reports?]]>
Author: jlegind@gbif.org
Created: 2016-06-29 14:28:54.747
Updated: 2016-06-29 14:28:54.747
The issue here is that ICIMOD is a regional initiative for the Hindu Kush & Himalayas that contains the members Afghanistan, Bangladesh, Bhutan, China, India, Myanmar, Nepal, and Pakistan.
I would rather not attribute all this data to a single country.
Author: kylecopas
Created: 2016-06-29 14:35:08.45
Updated: 2016-06-29 14:35:08.45
Okay, so just to be clear—by not including a country address to the organization, it's not counting any country's total.
I know it doesn't resolve our conundrum of reporting countries of publication and country locations for any given records...but wouldn't the same logic lead us to remove US from Cornell Lab of Ornithology, since eBird contains data from multiple countries?
Author: jlegind@gbif.org
Created: 2016-06-29 14:48:17.398
Updated: 2016-06-29 14:48:17.398
Yes - not stating the country is customary for international organizations.
The same logic could be applied to eBird, but not Cornell since they publish other resources. The eBird dataset is one of these cases where we don't have any established policy to resolve attribution in a graceful manner. Same goes for iNaturalist.
Author: trobertson@gbif.org
Created: 2016-06-29 14:49:33.135
Updated: 2016-06-29 14:49:58.075
I can't disagree to Kyle C's remark, but please aware that doing so would mean the drops would show in the analytics charts (e.g. country reports).
We do not go back and retrospectively fix endorsements etc in those charts - it would be a huge undertaking to do so.
Author: kylecopas
Created: 2016-06-29 14:57:01.129
Updated: 2016-06-29 14:57:01.129
Okay, thanks. I recognize that this is both difficult and messy. But, with no snark intended, here's another counter example of how inconsistently we apply these rules of thumb, fwiw. NatureServe is an international organization, but it would appear to we probably attribute its 165,736 Canadian records to the U.S, given the org's U.S. address (loving these new facets, btw!):
http://www.gbif.org/occurrence/search?datasetkey=7fd12114-9010-4c13-8f46-990fe04ca882&country=CA
I suspect this matters less where we don't notice these changes (like the US) as opposed to data-poor regions.
Note, too, by sheer dumb luck, NatureServe happens to have updated their records yesterday, which I think must be the source of yesterday's 1 million-plus spike in records. I will try to find out what's up—changes at the top a couple weeks ago, leaving my best friend there as interim CEO:
http://www.gbif.org/occurrence/search?datasetkey=7fd12114-9010-4c13-8f46-990fe04ca882