Incorporate Leading IDs into Top Identifier stats

Platform(s): website, e.g. Chrome browser.

URLs and screenshots: Delta Leaderboard, Hylaeus Leaderboard

Description of need and feature request details:

The Top Identifier stats on Observations, taxon pages, and Explore Leaderboards is currently calculated by the number of IDs of any kind (Leading, Improving, or Supporting). Many users become confused or complain if they feel the top identifier isn’t the one who knows the most species e.g. for a genus or family. Also, many observers copy and paste usernames from Top stats to ask for IDs, assuming the users listed are the top experts.

In the examples, I’m currently at the top rank for global Delta wasps and Hawaiian Hylaeus bees. The first example may be roughly accurate for which users know the most species, but the second isn’t. In part this is due to the stats title being somewhat misleading (Top Identifier), although it’s a catchy stats title. Something like Frequent Identifier or Most Identifications by may fit better. Regardless of the title (not the main focus), an improvement would be to incorporate number of IDs and number or percent of Leading IDs into Top stats.

Caveats:

  • The exact calculation to incorporate Leading IDs into Top stats could be discussed e.g. by developers or statisticians. I suggest incorporating total IDs and Leading IDs in a ratio of 70:30, or no more than 50:50.

  • This request assumes iNat will only use one Top stats, but would be compatible if multiple different stats were calculated and displayed separately.

  • This request would be compatible with additional ways to incorporate more into Top stats which could be discussed. For example, if Top stats (e.g. for a genus) also incorporated number or percent of finer IDs (e.g. of species IDs for the genus).

  • This request is compatible but not dependent on changing the Top stats title.

This is something I’ve been thinking about a lot recently as a way to reduce the incentive for people to just add lots of unfounded agreeing IDs because they want to be listed as the top identifier of something or somewhere.

I suspect a barrier to using an algorithm that combined multiple types of IDs would be added complexity and slower loading times. Using just improving or leading IDs would be simpler but might bias the listings towards people who have more time to spend on iNat and thus are able to add correct IDs first and it might make people rush when adding IDs so they can beat others and get more improving or leading IDs.

17 Likes

I don’t know the best way to do it, but I certainly agree something that refines identifier stats like this would be nice. I’m the “top” identifier in a lot of categories I know next to nothing about, just because I can ID things to that genus when I don’t know anything about the species level.

Geographic limitations on top identifier stats would be nice as well - I think I’m still the top IDer of Rubus globally, because I identified a lot of observations of the two most common species in my area.

That said I hope there’s still a way to look at the overall leaderboards, because although it’s not my motive for identifying, it’s still fun to be able to see the rankings and see how much I’ve done.

5 Likes

Maybe there could be a requirement to have made IDs on enough observations a certain distance apart to be featured in the global top identifiers list?

I feel like this would mostly just be an annoyance to the normal identifiers…

2 Likes

It’s an interesting idea but I don’t know if adding leading IDs would help in many cases. I’m the #2 identifier for Malvaceae but my IDs in that family are focused almost entirely on a single genus that I’m researching. I know very little about the rest of the family. Possibly a more useful number would be the number of taxa identified within the broader taxon, which would show a broader knowledge rather than a specialty in only a few taxa.

8 Likes

Might want to take into account the experts that don’t have time to go through observations with broad IDs to look for what they are an expert on. They might just go through and confirm species maybe also looking at observations that are already research grade.

I was the top identifier of Maryland Milkwort (Polygala mariana) and I got tagged by someone in Georgia. There are some lookalike species of Polygala but I only ID P. mariana here in east Texas where those other species aren’t found. So I couldn’t help the person at all.

2 Likes

I am always taken aback if I see my own name as a Top Identifier. Not me!

Frequent? Yes, definitely - I like to clear Unknowns. That is more about, retired with time, and a mindset for Unknowns.

When I look for an identifier - I decide which names I recognise, for their skill. Not sure that any automated algorithm would be useful to me.

For my Cape Peninsula default on Hylaeus https://www.inaturalist.org/observations?verifiable=true&taxon_id=118582&place_id=&preferred_place_id=1&locale=en&view=identifiers (Ignore number 16 - that was my 1 agreement)
It is easy for me to scan the list and decide who is worth asking - might they know? - do they reply? Or to tweak up taxon level or geography for more choices. Bonus points when iNat again brings me, on that obs, a new to me skilled identifier for just that!

2 Likes

As a person that does have quite a few ID’s, I would like to share how I approach it. Firstly, I do a lot of my ID’ing from the thumbnails, with ‘needs ID’ as a selected filter. So I often don’t know that several other have recently ID’d the same observation.

1 Like

Good point. I also didn’t even look at stats themselves my first few months IDing. I try to partially ignore stats in the sense of not letting them influence how I’d otherwise ID. Also, considering there are leaderboards for every combination of taxon and spatial scale, it would be hard for frequent identifiers to even know all their leaderboard ranks. About the request slowing site speed, in making this request or others I’m assuming iNat would upgrade, redesign, and/or reallocate resources (in some way) sufficient to allow implementing the features without slowing speed. Although it is good to keep in mind if a speed reduction may be a difficulty or trade-off involved.

Overall, in the event stats weren’t to change, it would help for iNat and the identifier community to do more to clarify common Top stats misconceptions to users. Top doesn’t always (although can) indicate best expert. And on the other hand, users who reach Top ranks typically haven’t done anything wrong (and if they had that’s supposed to be looked into separately, i.e. not to continue). Many users often complain if the best expert isn’t top identifier. But, anyone who takes themselves to be both things can simply make all the same IDs, and in doing so their greater expertise may also become evident. Given that Top doesn’t necessary mean best, that expectation and judgment should be shifted, and it seems unhelpful to overly worry and compete about stats. On the other hand, slight competition in a more healthy way can be a game-like driving force to spur on identifiers to try their best.

1 Like

I believe that the enthusiasm of being the first or of leveling up should not be pursued or frowned upon. Rather it could be redirected towards Taxa, Species, People and places that need someone competitive.

1 Like

Can’t edit my comment?
That link is for Delta wasp