ID A Thon Unofficial Guide

General
21

Thanks! I had thought that Unknowns were a minor part of the Needs ID pile, but I am wrong. If this link is correct - https://www.inaturalist.org/observations?iconic_taxa=unknown&identified=false&place_id=any&quality_grade=needs_id - Unknowns are more than one-third of Needs ID observations. That’s a lot! Guess I better target some of those starting Monday as well.

ETA: Whoops, my math skills are not functioning, apparently. I thought more than 500,000 Unknowns was a third of all Needs ID observations. Nope. Sorry about that! In fact, right now the 585,868 Unknowns are 0.0054% of the 108,406,822 Needs ID observations.

2 Likes
22

@lynnharper it’s actually 0.54%, or 1 in 185 Needs IDs are UNKs. Still a shockingly high percentage, where all someone has to do is add a general ID to not make it an unknown.

2 Likes
23

That’s what I’m planning! I’ll be working through as many spiders as I can!

6 Likes
24

Do you happen to know if there’s a simple way to filter out certain categories at a lower level?

For example, so I could look at all needs ID plants in my area except grasses?

It seems like there should be, but I’ve been unable to find such an option.

I understand that in the example above, I would still see any grass observations that were stuck at “plants” “monocots” et cetera (and hopefully at least be able to kick them to grasses) but it would help a lot to filter out those already marked at “grasses” or lower. Because if someone’s looking for confirmation on what type of grass, I have nothing to contribute there.

2 Likes
25

Oh, that’s funny - clearly, my math skills are shockingly low right now.

1 Like
26

Perhaps use this exclusion metthod? https://forum.inaturalist.org/t/how-to-use-inaturalists-search-urls-wiki-part-2-of-2/18792

Try https://www.inaturalist.org/observations/identify?verifiable=true&place_id=any&without_taxon_id=47434 .

3 Likes
27

Even if you have no idea what the observation is, please add some annotations if you can so that the specialist IDer can find those observations eg eggs, Larvae, tracks, scat, juvenile.

8 Likes
28

While getting “unknowns” out of the pit of doom is helpful, I personally just want people find a strategy that works for them, and that hopefully they can continue and grow in using beyond the ID a thon. That includes folks who know grasses backwards and front, or are really good at identifying birds in out of focus or faraway shots. And folks who are just trying to find a way to help.

If folks don’t know how to get “unknowns” filtered in such a way that they aren’t staring at a lot of microbes or kingdom disagreements, I suspect they will get frustrated quickly. I’d rather have someone who is really good with angiosperms zip through those classifications than get frustrated and quit and not feel like their contributions are worthwhile. One of my coarse ID defaults when working with “unknowns” is dumping them into angiosperms (assuming of course they are one), so someone else working on that classification helps out too.

I also want to note a few things with unknowns, as I’ve been focused on them for a few weeks.

(1) the oldest “unknowns” often have some problem with them that the observer hasn’t resolved or responded to. These include duplicate observations, landscape photos, observations with more than one organism, or taxon set to the observer’s selection, and not the community taxon (meaning that no matter how many IDs there are, the observation stays parked as “unknown”. A corollary of this are observations that just simply can’t be ID’d, like someone’s 1x1 pipe square photos of the ground (there’s a name for these things and I can’t remember it bc my memory is crap lately).

(2) sometimes an observer will put a species or taxon suggestion in the comments - use this as a guide to help you when reviewing computer vision suggestions. Just today I was staring at a photo trying to figure out what I was looking at, and then finally I actually paid attention to the observer’s note - “nudibranch”.
(a) Sometimes the observer’s notes are in another language - Google translate is your friend. (b) Sometimes the observer’s notes have a common name for the plant that, while written out in the Latin alphabet, is a common name used in the observer’s country but perhaps is not something you are familiar with. Again, a quick internet search is your friend here.

(3) for reasons that are above my mental knowledge pay grade, some algae are classified in the Plant kingdom, and some algae are in a different kingdom altogether. Yay taxonomy!

(4) coarse IDs are your friend, as is looking at the taxonomy trees of the computer vision suggestions. Are the suggestions all in the bean family and you’re having trouble narrowing it down? Use the coarser ID for Fabacae or the like. Every suggestion is a grass? Poaceae that observation. (Yes I am probably butchering the spelling sorry)

I probably have more thoughts, but that’s it for now.

7 Likes
29

This!!! Thank you for bringing this up. I do this when working with unknowns and it’s a really easy thing to do while you’re in the observation.

30

We have a (newish) DQA for that which pushes the obs to Casual. I also leave a copypasta comment - hoping the observer will respond, but at least warning the next identifier.

I also leave ‘opted out’ comments.

but first you have to help the ID to where iNat will allow that annotation. Cannot ‘flower’ unless the CID is at angiosperm first.

4 Likes
31

If you add &identified=false to a URL, it will remove everything except what I call true Unknowns. It gets rid of the bacteria, protists, and Life observaions.

6 Likes
32

That sort of misses the point, which is that the change from Unknown to something is a larger change than from genus to species when it comes to classifications and the usefulness of the database as a whole.

1 Like
33

You might see that differently if you made more IDs. In theory it may be a ‘larger’ change. In practice, dumped where it is ignored, is less about useful than knocking down the ‘vanity’ number of Unknowns. At least it has An ID. Yeah, thanks, not really.

Have been thru the GSB residue for South Africa, down to and including Family. Over to the taxon specialists and I can turn to the broad GSB IDs for the Rest of Africa.

iNat has a substantial identifier gap between people who catch today’s Unknowns and specialists who look only at their taxon.

2 Likes
34

Compared to getting to a species, I don’t think it is particularly useful to go from Unknown to Something if that Something sticks at Vertebrata or Magnoliopsida. There are not many places on earth where it is news that plants grow there. I am not ignoring the principle that getting to Magnoliopsida is a first step, just that getting there is not a great contribution to “the usefulness of the database as a whole”.

6 Likes
35

When I go through “Unknowns” there are a lot of plants I can identify to species languishing in there, and a fair number of both plants and fungi I can identify to genus, so it’s not just about sorting to kingdom. There’s something uniquely satisfying about finding species I instantly recognize that have waited a long time for an ID.

7 Likes
36

Another tip:

Find genera stuck at Needs ID and at genus level that you can ID confidently. Many genera are monotypic or have a few species that are easily distinguishable from each other.

6 Likes
37

Best thing ever when ID’g unknowns is when everything is the same family or genus, or there are less than 3 options on the computer vision. THE BEST

3 Likes
38

But the steps are qualitatively different. Sorting unknowns into iconic categories is far less cogitively demanding than distinguishing species within a genera (some of which can contain hundreds or even thousands of species). This is borne out by the fact that, generally speaking, identification keys become ever longer and more complex as you descend through the taxonomic tree. At each level, the cognitive load becomes progressively greater.

At present, our community does not have anywhere near the capacity to handle the collective cognitive load. And it’s not simply a question of numbers. Bringing in a thousand extra identifiers isn’t going to help much, if all they ever did was sort unknowns into iconic categories. All this does is push the problem onto several much larger piles.

The only way to increase our capacity to handle the load is to keep challenging ourselves to learn new stuff. This seems to be the primary focus of the ID-A-thon. Dwelling on the minutiae of personal priorities and productivity stats seems somewhat misguided. If we all just stick to our normal routines, very little is going to change in just one month.

9 Likes
39

You make a very good point here, @bazwal. Right this minute, there are 108,432,199 Needs ID observations worldwide. Can we reduce that by 1% in a month? Assuming my math is correct - and I’ve proven in this thread that you all should check my math! - that’s 1,084,322 IDs. Plus enough to cover however many new observations are added in that month, on top. Can we end up with only 107,347,877 Needs ID observations by the end of January 15th?

Boy, that’s an ambitious goal, to my mind. Does the number of Needs ID observations ever actually go down? Is the most we can hope for is to keep that number level over the month?

Or can we collectively change our efforts to make significantly more IDs than are usually made in a month? Is there any way to measure how many IDs were made from, say, November 15 to December 15 of this year?

Or is all this talk of numerical goals off-putting to people? Shall we simply have fun collectively identifying for the next month?

7 Likes
40

I am new to Identifying and I tried to look at Unknown observations in my area. I was frustrated because I wasn’t seeing the AI suggestions. It turned out there is a setting under the Suggestions tab for Source: that was defaulted to ‘Observations’. Once I changed this to ‘Visually Similar’, I was much more successful. Something to point out to Identifying newbies.

IMG_9996
IMG_9996569×953 121 KB

10 Likes