I have been matching iNaturalist place ids with Wikidata using OpenRefine, but get the id to place name mappings using the API. I am getting close to where I would consider doing this data scraping. Basically, I took a place on an iNaturalist observation and then iterate from x below and y above to that place id, where all ids between x and y are pointing to places in a single country. I then let OpenRefine do the reconciliation.
This works quite well (for those places that do exist on Wikidata and I do not create entities for those that do not match), so if a zip file is available, I could continue doing this, without having to hamper the API.
Sorry, the cron job we had to generate the file was referring to a hard-coded db host name and we changed our db host in early Feb 2021, so the cron job broke. I regenerated the file and am updating the cron job so this wonât happen again.
Thank you, I hope the 120.210 Places (nov2020) gets an update but no idea how easy it is to match. It seems Western Europe is not Ready yet.
2894 resultaten in 1170 m
4752 resultaten in 5711 ms (2 weeks)
( Open Refine. Zie: https://w.wiki/37MB
Germany heeft iemand hier boven gedaan en Andra heeft Suriname en Mongolia gedaan Open Refine. Zie: https://w.wiki/37MB het amazonewoud But it should be automated by some way with a track and trace or a logging but i have no idea how to write an bot or how to automate it. I am afraid someone who know how it works should do it. Maybe it was CHRISTIANSW https://forum.inaturalist.org/t/use-wikidata-for-place-names-and-wikipedia-descriptions/7702/2
Map of places i/37Sx (Query)
3369 resultaten in 6743 ms
4594 resultaten in 8628 ms (Tuesday 23mrt)
4752 resultaten in 5711 ms (2 weeks)
Mix en Match https://mix-n-match.toolforge.org/#/catalog/2738 VerspreidingsAtlas.nl
Thank you. For me MixandMatch and OPEN REFINE is just complicated, no idea how it works.
3869 resultaten in 173 ms
Monday 4594 resultaten in 6108 ms
4752 resultaten in 5711 ms (2 weeks) Uploading: image.pngâŚ
Is there some one who can give the translation of PLACE_TYPE in the file ?
6 could be a street or a place
9 Place Dorp, City, Stad
100 could be âNature Reserveâ
33 could be âEstateâ
Place_Type =7,20,33,100,101,1011
And is there a difference between Open Refine and Mix And Match?
Type of place to retrieve
Allowed values: Aggregate | Airport | Colloquial | Continent | Country | County | Drainage | Estate | Historical County | Historical State | Historical Town | Intersection | Island | Land Feature | Local Administrative Area | Miscellaneous | Nationality | Open Space | Point of Interest | Region | State | Street | Street | Street Segment | Supername | Town | Undefined | Zone
https://forum.inaturalist.org/t/is-there-a-place-where-i-can-go-to-get-a-list-of-inaturalist-place-ids/4016/3 ou can also get place_id lists using the search URLs, as described here: https://forum.inaturalist.org/t/how-to-use-inaturalists-search-urls-wiki 21.
For example, the following URL returns JSON data for all the places in Oregon (ancestor_id=10) that are open space (place_type=open+space) and have the word âParkâ in their description (q=Park):
Each JSON block starts with the place_id and the place name. Unfortunately, you seem to be interested in California (ancestor_id=14), which has a LOT more places defined than Oregon. The JSON interface limits you to only 200 items per search (per_page=200 max). To go beyond that, you need to get the first page, save it, then increment the page counter (page=1, page=2, page=3, âŚ) until one returns fewer than 200 items.
The complication is not so much with the tools, but more with the data. The most challenging thing is to reconciliate the place names with what the items in Wikidata. For example, in the Netherlands it is not easy to distinguish a place, a municipality, a province or a recombined municipality. I saw the same complexity in Peru, where names are even shared by rivers, places and provinces.
At first I thought the latitude and longitude could help, but then I ran into places like âTexelâ which is both a municipality and an island and âSchiermonnikoogâ which in wikidata shares both the status of a municipality and and island.
Based on the data I got it is not easy to distinguish what type of place names we are dealing with. I suspect that the place-type could help here, but I have yet to understand to what the numbers relate. Is there a data dictionary about the place name table?
Once identification of the place type is possible and combining that with latitude and longitude, adding the place ids should be straightforward.
I must admit that it was fun trying to add place ids to Wikidata using open refine. I had to get local awereness to make the right choices which felt a bit like a virtual trip to the region. As you might notice some countries like chile, panama, mongolia, etc are better covered than other. I virtually travelled to those regions :D. The downside is that this approach does not scale.
If it is a bad idea to place the video one just can take it away without any problems
Wikipedia Weekly Network - Live editing Wikipedia: Biodiversity edition #3#EarthDay Honoring #EarthDayâ, Users Siobhan Leachman, Andra Waagmeester, Mike Dickison, Albin Larsson and Jan Ainali are joined by iNaturalist co-founder Ken Ichi to discuss and showcase how iNaturalist is used in Wikimedia. Weâll start with the new property iNaturalist place ID in Wikidata and see where our discussion leads us. Join in the chat to ask questions during the session.
https://en.wikipedia.org/wiki/WOEID A WOEID (Where On Earth IDentifier) is a unique 32-bit reference identifier, originally defined by GeoPlanet and now assigned by Yahoo!, that identifies any feature on Earth.
If anyone is still looking for this, itâs here, and admin levels are here. Note that while pretty much all places at the continent, country, state, and county levels will have their admin_level set correctly, many towns and parks will not have admin_levels.
Can you give an idea of how many iNat places have a woeid set? Spot-checking suggests itâs not all that many, but if Iâm wrong, that might be nice to have in the export.
About 5174 places have a WOEID set. FWIW, these were mostly originally imported from Yahooâs now defunct GeoPlanet API and subsequently assigned a boundary, mostly by users (I assume). Iâll include them in the export, with the caveat that I have no idea if theyâre accurate.
