Link to top identifiers of that taxon, not *observations* of that taxon

Feature Requests
#15

In two sentences, the taxon page-associated, identifiers’ leaderboard that is pointed by the yellow arrow above should generate the following table :
https://jumear.github.io/stirfry/iNatAPIv1_identifications_identifiers.html?taxon_id=98413&own_observation=false (notably its columns “rank”, “login id”, “ids”) (example for taxon 98413)

instead of the table currently associated with it :
https://www.inaturalist.org/observations?taxon_id=98413&view=identifiers

Is that possible ?

This is my first question.
A second question I would like to ask, as a possible improvement, is best explained in the very next post #16.
( for the sake of simplicity I have removed the first variant of it, which was included in this post ; indeed it introduced a somewhat too visible and debatable modification of the taxa web page).

#16

This is my second question, based on another example (screenshot) :

image
image938×388 75.9 KB

I would propose to add an “overlays” icon next to the “top identifier” (or any other icon that would allow to click options), as shown in the following maquette :

image
image843×693 89.5 KB

Clicking on that icon would show what follows :

image
image938×460 91.9 KB

Indeed this is the default way of counting IDs (see above).
But now we are allowed to choose options.
And we could click “own observations” and unclick “supporting”, thus reflecting @bouteloua @joe_fish and my suggestions here and other posts, that is :
image
image942×460 91.4 KB

which leads to a different top identifier, which will appear as follows, after we move the cursor away from the overlays window so that the latter disappears :
image
image942×460 80.4 KB

As explained in the previous post, the page now shows the top identifier taken from this table :
image
whereas the current top identifier definition, as represented in the first, unmodified screenshot, comes from this table :
image

In this case, one can see that there is absolutely no overlap between the 5 top identifiers as currently defined and the 5 top identifiers defined by the leading/improving categories including own observations. And anyone could choose his/her way of defining a top identifier.

Of course, the leaderboard would open the corresponding tables.

I hope this proposal is an … improvement.

2 Likes
Crazy prolific but incorrect identifier
#17

This display (whatever it could actually mean) is counterintuitive and incomprehensible (in short, it is inconsistent, forgive me):

image
image1176×274 81.7 KB

image
image1192×808 63.1 KB

image
image1199×348 104 KB

image
image1194×816 58.7 KB

2 Likes
#18

Another related thread:
https://forum.inaturalist.org/t/only-1-identifier-for-species-but-top-identifier-is-someone-else/11177

#19

This is my question 1 (post #15), which (as far as I understand !) is the same as @bouteloua’s initial question for this topic. Presently indeed, if you click the leaderboard under the “top identifier” on the https://www.inaturalist.org/taxa/ page (like your https://www.inaturalist.org/taxa/62900-Brugmansia) your are led to the kind of results that you just found :
https://www.inaturalist.org/observations?taxon_id=…
which usually gives a different top identifier (and inconsistent leaderboard) compared to the taxa page for the reasons discussed in the very first posts : yes, the taxa page identifiers are differnent from the observations page identifiers and @bouteloua started this topic here to ask to correct for this inconsistency.

To do so, it helps to realize that the “top identifier” of the taxa page, along with the full corresponding leaderboard, can be obtained with this kind of url :
https://jumear.github.io/stirfry/iNatAPIv1_identifications_identifiers.html?taxon_id=62900&own_observation=false
You will find here that the top identifier is the same as shown in the taxa page.

The leaderboard icon from the taxa page should not link to the observations page associated leaderboard. In fact (see above) in the observations page the so-called identifiers should rather be termed something like “all people involved in the identification process” (with correct or incorrect, vague or precise IDs).

It is critical to realize that there is such an inconsistency, as you also do, and the first step is to correct it.

#20

Yes, this explains what we see.

Functionaly, I don’t need (do you?) to know the people [most] involved in the whole identification process. I need to know how has proposed this taxon (or any other one at a lower rank) as an ID. This is the initial feature request, if I understand correctly.

As an inconsistency with the linked leaderboard, I would consider it also as a bug.

Because all information needed is already available (already computed for the leaderboard page), and because the fix consists “only” in changing an identifier and a count in the Taxon page, I don’t understand why it is not fixed after a year.

Of course it might not be prioritary, but as a bug I suggest that it need no further discussion (or votes).

Maybe it is not fixed yet because this change request is registered as a feature request, instead of a bug.

1 Like
#21

A bug or winged poetry…, I do not know. But once we agree on this fundamental question, not only should it be implemented, but we could discuss on possible improvements : notably giving the choice to select among the main different types of identifications and whether they should include one’s own observations (my question 2, post #16)…

#22

It is a real improvement for finding people that are likely able to help identifying/reviewing new observations.

#23

Thank you. As briefly mentioned above, several different posts (two in this topic, #4 and #5 and quite many others) make essentially the same suggestion.
The only addition here is the checkbox structure (opened by an overlays icon or anything more appropriate) on the taxa page, thereby introducing flexibility.

Some additional references :
https://forum.inaturalist.org/t/gamify-accuracy-award-value-to-quality-not-just-quantity/14428/36
@bouteloua "One idea I had fleshed out, but don’t think I ever posted, was to show only improving, or only improving+leading IDs in the default view of stats. i.e. don’t tally all the supporting IDs. "

https://forum.inaturalist.org/t/unknowledgeable-commentators/15441/33
@bazwal “The lists of top identifiers for taxa would be a lot more useful if they took into account the proportion of improving IDs. If someone at the top of the leaderboard has 95% agreeing IDs, that doesn’t tell you much about how knowledgeable they are.”

#24

@bouteloua proposed a specification a bit more complete:
https://forum.inaturalist.org/t/gamify-accuracy-award-value-to-quality-not-just-quantity/14428/36

With a checkbox for one more identification type: “Maverick”.
With separate checkboxes “For others” and “For self”.

#25

Would an additional checkbox “Lower ranks only” be useful?

2 similar use cases:

When reviewing observations identified at the rank Kingdom Plantae, when I see a Family Crassulaceae observation, I simply identify it at the Family rank (because I have almost no knowledge in this Family). Now, suppose I make numerous such identifications, I would then appear in the leaderboard of the Family. But in general I will not be able to help identifying Crassulaceae observations at a lower rank, so it’s better to forget about me for Crassulaceae, and a checkbox “Lower ranks only” would ensure that.

I also made many identifications at the rank Subfamily Faboideae, in most cases because most people ID at the rank Family Fabaceae instead, and also because in many cases I can’t be sure of the Tribe (in general I also add a comment to propose one or a few possible Tribes, which may help other people to further investigate). These many identifications at the rank Subfamily may be useful, but a checkbox “Lower ranks only” would preferably ignore them.

(Someone in South Africa complained about these many identifications at the rank Subfamily Faboideae, saying it’s pointless. I answered that Caesalpinioideae and Mimosoideae observations would be much easier to find if so many observations of Faboideae were not left at the Family rank. Of course, if you care only about Faboideae, you don’t mind if they are all identified at the Family rank by default).

#26

Yes, maverick is the fourth category in addition to supporting, leading, improving. I just doubt that it could be helpful because a maverick can be right and equivalent to a “leading” yet against the community. But if that “maverick” is right, the community will quickly take side with her/him, and the category will quickly switch to “improving” - whereas the “leading” category is interesting when there is no or only a small community and nobody around that would be able to provide “supporting” IDs !

As for “for self” vs “for others”, this is what I meant with a single checkbox for “own observations” (corresponding to that url option “own_observation=true” vs “=false”).

#27

Concerning your “Lower ranks only” proposal, to me it is already included in the identifier definition that we are calling for here, as this identifier definition refers to any given ID per se - thus higher levels are excluded, and lower levels are included by default.

#28

The current taxon filter definition is “this exact taxon (at the rank of this taxon), or any other lower rank taxon (still matching this taxon)”.

A “Lower ranks only” option would switch the definition to “not this exact taxon (at the rank of this taxon), but any lower rank taxon (still matching this taxon)”. It is not included in the current definition, because without this extra option there is presently no way to obtain the same leaderboard results in the iNaturalist web application.

The idea is to find people with knowlege within, say, a Family, not people that are only able to say if a specimen belongs or not to the Family. The former is the common use case, for improving an identification. The later is rare, I mean having identified a Family, without being sure of it at all, and thus not yet caring about improving the identification at a rank lower than the Family.

1 Like
#29

I understand now, thank you - then 5 checkboxes for the identifiers leaderboard : improving, leading, supporting, “own observations”, and “lower ranks only”.

Perhaps @pisum could kindly tell us if this or part of this is feasible ?

Again, this would be in addition to point 1, that is correcting the identifiers leaderboard (notably to make it consistent with the top identifier shown in the taxa page).

1 Like
#30

Provided that “Maverick” is automatically included when selecting “Leading”, as you explained above.

#31

i’m just a regular user of the system, not staff. so i don’t really have insight into what staff would be willing to change. personally, i think changing the Taxon screen as you’re proposing makes that screen too complicated.

and before you even think about changes to the Taxon screen, you really need to address the fundamental difference between identifiers of observations and identifiers (of taxa). there are different reasons you might want to get each of these. so i would leave the Identifiers tab on the Observations screen alone.

in my opinion, the right way to address this feature request is to create a new Identifications screen that is similar to the Observations screen. instead of allowing you to query for observations and related statistics, an Identifications screen would allow you to query for identifications and related statistics.

once an Identifications screen exists, then you could link the top identifier in the Taxon screen to the top Identifiers tab in the Identifications screen.

doing things the way i’ve suggested above is not a small thing, which is probably why no action has been taken down this path. still, ultimately, i think going down that path is needed as part of raising the importance of Identifying to a level more equal to that of Observing.

1 Like
#32

Let’s be interested in the taxon page of Family Apiaceae and let’s consider an observation of a Wild Carrot. Someone that first identified this observation as “Plantae” is one of the “identifiers of observations”. Someone that later identified this observation as “Apiaceae” or as “Wild Carrot” is one of the “identifiers of observations” and is one of the “identifiers (of taxa)”. Did I understand correctly?

Could you elaborate? What are the possible reasons to get the “identifiers of observations”?

(I don’t question the existence of the need, I would like to get the use cases and user stories that the solution depends on).

In the case the observer did not add his own ID on an observation, the person who identifies this observation as “Plantae” do this in order to put the observation into a category so that it may be found by identifiers interested in “Plantae”. In this case, getting the “identifiers of observations” would mean getting these people who identify as “Plantae” in order to direct observations to other identifiers. Is that the subject?

#33

yes.

there are times when you’re not specifying a particular taxon or rank but you still want to know who’s helped to identify. for example, if you’re running a bioblitz and are interested in seeing not just who’s observing but who’s helping to identify (any taxon), then you’d be more interested in seeing identifiers at an observation level rather than identifiers at the level of a particular taxon.

this applies even if you, say, had a taxon-specific Daucus carota project. you might still want to see everyone who helped to identify observations in that project, even if some of those people could only identify to Apiaceae or Plantae.

1 Like
#34

@pisum : the initial request in this topic was for a leaderboard for identifiers in the taxa page where the identifiers would be the identifiers of the corresponding taxon, thus “identifiers of taxa” as you call them.
One year later we notice that apparently this has become true for the “top identifier” shown in the taxa page but still not for the leaderboard, which still links to the observations page.

So there is a problem of consistency in the taxa page, in addition to the problem of definition of the identifiers.

And the problem of definition deserves to be addressed more precisely than just using your least stringent definition, where anyone who added any kind of ID, including completely wrong IDs, would be regarded just like the best available expert in taxonomy.

A qualifying name should designate them not as “identifiers” but something like “incomplete identifiers” (if we neglect the wrong ones).

We could add a 6th checkbox to the proposal , possibly termed “incomplete IDers included” (yes or no). Here are the 5 others :