Some time back, I posted this question ( https://forum.inaturalist.org/t/why-getting-palaearctic-species-suggestions-for-nearctic-observations/48165 ) about why I keep finding Nearctic Donaciinae beetles identified as European species. Two answers emerged from doing the corrections and 10,000+ identifications of these beetles: 1) Originally, almost no donaciines were correcly identified in North America so naturally all new observtions only were given such suggestions. Over a period of 6 months or so I made all the corrections and the frequeny of european misidentifications diminished as folks went in and Agreed my IDs once they discovered that I know these beetles pretty well. Yipee! I now have followers and some folks who are out chasing donaciines because they can get correct names on them! When someone wants to select “Suggestions” or “Compare” in the ID process, rarely or no european taxa present as choices. Mission accomplished.
2) However, where they DO come up is during the uploading of new observations (or when making identifications) and you type the name beginning with genus. Typing in “Plateumaris… “ the top suggestion remains Plateumaris sericea, a european species with 2226 observations, almost half the total observations for the genus. So the occasional european identification pops up but I fix that… Maybe if the Wikipedia pages were all written for donaciines people could read up on them before picking a name! Another task I’ll need to learn.
So the question begs, why does iNat do this at all any longer, now that the Nearctic donaciines are cleaned up? Just curious…
3 Likes
When you start typing a name, the taxon list is not populated based on location or visual similarity. It fills based on a string match to the letters you type, sorted by number of observations.
6 Likes
OK thanks. I thought there was some sort of algorithmic logic there but I couldn’t put it to words. But shouldn’t it work based on loation to some degree? Isn’t that what the “Suggestion” and “Compare” things do? Those are clearly location-based because they still come up empty or short when there are no iNat data in the area.
1 Like
If you mean “wouldn’t it work better if it were biased based on location”, then I think the answer is “probably”. If you mean “isn’t it a bug that it’s not biased based on location”, then I think the answer is “no, it was never designed to use location”.
I agree it’s also maybe a confusing choice for the UI to work that way – if you just click in the box and wait, you get a suggestion based on visual similarity and location, but if you start typing, visual similarity and location no longer matter.
I guess I’m not sure how I would change it to make it more obvious though.
4 Likes
I think the assumption is that, if you’re using the CV, you select from one of the suggestions provided. If you begin typing in a name, you are not using the CV.
1 Like
I’ll also add a link to the tangentially related request for better sorting of names when you start typing in the taxon field:
https://forum.inaturalist.org/t/prioritize-scientific-name-over-synonym-and-names-in-other-languages-in-species-search/23191
5 Likes
If you start with the broad ID you are certain of, and you have added the location - you should get a good (or at least better) ID. You can toggle to Visually Similar - then you forfeit location - works better for Cultivated / Captive or Invasives. Text will pull from scientific names, and also common names in any language for which they have been added to iNat.
1 Like
