Show people what to do about a missing taxon in taxon search

Feature Requests
,
1

Platform(s), such as mobile, website, API, other:
Mobile and Website

URLs (aka web addresses) of any pages, if relevant:
all observation pages

Description of need:*
I have seen dozens of times a new researcher or identifier in comments not adding a confident taxon level ID simply because they find it in missing in inat and don’t know what to do.
*
Feature request details:
In suggestions indicator, when a full species name is missing from taxon, it would be very helpful to pop a link like “Learn how to flag new taxon on iNaturalist” and then include a link to page with screenshots and what to do properly when flagging. i know it’s a simple change but it will actually directly help the people who search the full name and don’t know what to do in those cases (i know the help pages exist, but often they may not go and search those things).

I have personally seen people ask or clueless that they can raise flags (although i understand not everyone should raise flags for every synonym or mistaken things) the positive side of people able to learn they can add species into and how to do with such direct indication at most possible place - is right below the name suggestion - as below screenshot from identification bar when typed manually such names fully. this can be implemented everywhere and hopefully will aid finding more exact species.

ps: search external providers never works whenever i do, maybe its broken or idk.

the only BIG caveat would be implementation details of such feature and how long before should the autosuggest bail up and do this indicator. and sometimes people may simple be misspelling (Also why isnt simple levenshtein distance based typo indicator (like how google search already does did you mean) when nothing matches search already considered for this search bar? i think that would aid a lot too)

image
image690×308 22.7 KB

2

I take their typo to Google, come back and check iNat, then flag for missing sp.

Placeholder because typo / missing sp is why we made that project

https://www.inaturalist.org/projects/placeholder-backup/journal/98782-placeholder-backup-project-what-is-it-and-why

Making it easier for iNatters to do that directly would be good. At the flag (for missing sp) they need to be prompted to add the source (POWO for plants) and the obs where the sp is needed.

2 Likes
3

I like the idea of dealing with typos better to get to taxa, though with all the permutations of different taxon names on iNat (scientific, common, languages), it might be hard to strike a good balance with implementing as opposed to suggesting variants - this could actually make it harder to find the intended taxon in some cases if a lot of less taxonomically relevant matches are suggested. But if no other matches come up at all, checking for typos could be helpful.

Honestly, I would be in favor of removing the “Search external name providers” link since taxa added this way cause a good chunk of taxonomic issues/poorly added taxa at this point.

5 Likes
4

sometimes i see google bails too for arcane species, google’s autocorrect is focused on their frequent search queries which is understandable from their design point of speed and not making too many ad-hoc suggestions. thanks for project link.

but a separate typo check system for inaturalist can better deal this with more possible taxa we already have here. (maybe the typo correcter can also take feedback from geomodel score) - uhh now idk if these two things i requested should be separated as two requests, they are connected though.

5

yes i agree the system design is bit tricky in striking balance.

it can be combined with vision/geomodel suggestions to strike better suggestion maybe only when there are multiple collisions for that particular typo (again its very tricky to estimate when exactly should the typo correcter trigger over conventional matches in this case)

it can even be given manual trigger for this case (like how current AI suggestions not from area can be triggered explicitly in that bar) with list of all likely name collisions. these typo results can also show geomodel/vision tag below it from those scores - nearby/visually similar

But if no other matches come up at all, checking for typos could be helpful.

exactly. if there is only one obvious typo species (edit distance score of 1 or 2) in entire iNat list and zero true matches to anything they typed, then the iNat popping it will be really helpful there and its really easy to implement such case especially as there is no real autocorrect for all species names elsewhere. (i can contribute technically and draft codes too if this is considered)

With clear indication of something like: Did you mean?:

Honestly, I would be in favor of removing the “Search external name providers” link since taxa added this way cause a good chunk of taxonomic issues/poorly added taxa at this point.

i heard in a flag comment that it is using old taxonomic export from COL 2012 data or so and so some autoimports are gonna issues from that data especially with changes in names or synonymy (maybe they were cautious to not update it completely if it collides with current trees or not remove it fully to reduce curator work?)

—

ofc there is also phonetic Double Metaphone and better Jaro-winkler distance matching that can also be triggered further if everything in simple typo correct and conventional matching fails too.

Jaro-winkler is perfect one for those tricky cases of people finding almost everything correct in prefixes but mistyped some epithet ends which could have been matched with existing gender changed epithets on iNat from that.

—

I also believe there could be some Latin/Greek rules somewhere that such typo correcter can readily and simply combined with for reducing this autoimport issues of epithet gender based changes.
the correcter can even show very simple reasoning text below on why it is better match for those cases - for example this was recently auto-imported https://www.inaturalist.org/taxa/1647881-Eurybrachys-dotata
and it could have been avoided as there is already E. dotatus on iNat and such typo correcter system as above with simple grammar ending rules could have caught and popped a suggestion such as:
”Did you mean? : Eurybrachys dotatus
_in small font text under that:_ “dotatus is grammatical agreement of masculine Eurybrachys” (which can be generated from those latin/greek rules dataset in typo correcter) - Again i am not saying iNat should autocorrect all existing taxon genders without some curatorial work, but only when typo correcter duty it is to pop suggestions for existing species list matches on iNat

or if everything failed in such matchings, then popping direct indicative link in that popup place to a page showing how to raise flag and all details clearly for newbie.

1 Like
6

we do agree on one point that if there is no true matches the current autocomplete should trigger some typo-correcter or at bare minimum implementing such typo correcter designed explicitly for iNat names that can be triggered even manually is better than no such system or autocomplete as is now. personally i can sometimes recall jumbled name but atleast a close one that can be fuzzy matched as in this case (and probably there can be similar experiences for others too) - ofc one could comment a person not able to fully remember taxon name is not qualified to id that taxon but i can counter that with thousands of taxa, the mind can lag a bit to catch onto it until repeated IDs of it and if those are rare names especially - which again supports such matching system is better aligned with iNat principles to cover those critical unpopular taxa too in particular.