I don’t know if it would be feasible, but… @jeanphilippeb , it is possible to include in yellow label projects not only “unknown” observations, but also observations with coarse ID (coarser or equally coarse than the project taxon)?
This will solve the problem that @DianaStuder explained, that when someone move an obs from “unknown” to “Plants” the obs would not be included in yellow label projects, and thus not seen by IDers that use them.
Because that’s a research project all of its own (that’s also using inat data, and couldn’t get enough data to train its models with without something like it).
There’s nothing wrong with giving people an instant gratification best guess at what the thing they saw was, which they otherwise might have no idea at all where to start looking to figure that out. If that’s a gateway to them being hungry to learn more, that’s an awesome outcome.
It’s a tool. Right now it’s still a bit flint-knapped and attached to a handle with catgut and resin, but in the hands of good toolmakers it will get more refined over time.
Making that kind of distinction, with an implicit assumption that anything not produced by “professionals” is inferior and suspect is surely part of the problem.
Do you (being the collective rhetorical “you” that actually means ‘all of us’, not the fingerpointy one) really think that data produced by this year’s grad student who has no idea if they will still “have a job” in the same field in a couple of years time - or by someone paid minimum wage and given minimum training to collect it for A Professional - both with minimal supervison, has some assurance of being qualitatively better than a collection of data which has been extensively and openly peer reviewed over time spans of years?
Or that research performed by Professionals with competing time demands - to publish or perish and attend faculty meetings and lecture disinterested undergraduates, and up to their necks in Career Developing Politics and grant applications etc. etc. - is somehow superior to that of the people not trapped in that circle of hell, who simply have a genuine interest and obsession in learning the actual truth, not just what conclusions will kick you up the most cited leaderboards …
Sloppy research is sloppy research - and the only thing that makes Professionals seem less prone to it is an institutional culture that makes pointing it out a generally greater and more career killing sin than performing it.
Go confident and withdraw promptly.
The worst ID is an abandoned WRONG one, due to that more than 2/3 CID algorithm. If 2 agree on the wrong ID we need another 5 to counter them, instead of just 2 to agree.
If you are not happy to have someone agree in support of your ID, then your gut is telling you to make the ID broader - where you are comfortable with support. I see IDs as gridlock in a busy city - if I can move that one, that way, next.
Blame some of that on iNat conflating Needs ID and Not Wild.
People want an ID, so they are ‘forced’ to pretend the obs is Wild so it stays in the Needs ID pool.
If Wild or Not Wild was separate from Needs ID or CID - we wouldn’t have a logic fail we need to work around.
(I do mark my Not Wild with iNat’s confusing jargon of Captive / Cultivated)
Let us crunch the numbers.
Honest Unknowns are a large and highly visible issue, that many people work on. We can filter by location.
600 K
Greenwashed to Plant Kingdom
840 K
Since broad plant IDs above family are also in limbo, we will use that as a more realistic target
3.4 million obs
How is moving an obs from a visible half million, to a binned three and a half million better? How do we solve my 3.5 million problem ?? Enter @jeanphilippeb 's projects with their yellow and blue labels.
Tweak my links to the location where you keep up with IDs to see your own local issue, where you can work with your ID team. Dream on Diana.
@jeanphilippeb missed our 24/7 ?
Surely this can also help hint the CV to improve the accuracy of those generated project lists. There’s been plenty of discussion about how the CV can make bad decisions on blind images, but does much better if it’s hinted toward the correct branch of the tree.
There were definitely things that weren’t birds in the page of probably-birds he linked which the CV wouldn’t have called birds if they’d been (correctly) “greenwashed” first.
I’m not saying everyone should race out and tag every unknown as “Plant” or “Animal” (which is about the same problem) - but it at least puts them in the right pile for the people who mostly specialise in one or the other (and not being confident to ID many plants doesn’t mean we don’t love the people who are!) and gives the CV its best chance to not be Horribly Wrong.
Offering an ID that’s much more specific than Kingdom clearly has the most benefit, but one that’s correct but less specific should never hurt. If it does that’s a symptom we should see a doctor about.
Just wait a couple of years until “Honest Unknowns” hits that watermark then think back to the Good Old Days when Broader IDs were ‘only that many’ ? :D
How long do the projection of current stats say we need to wait for that to happen?
We need more identifiers who are more specific generalists (people who go through “Plantae” or “Magnoliopsida”, “Animalia”, “Fungi” etc.). But I don’t agree with letting observations sit in “unknown” if someone can ID them to Kingdom.
If we leave all those observations in unknown, they will drown out the true unknowns, taxa that are often even more neglected than (vascular) plants (which incidentally are not that neglected in my area to begin with).
Another reason to check “Plantae”, “Magnoliopsida”, or “Liliopsida” anyway, rather than just looking at "Unknown"s, are disagreements. If someone posts a wrong CV ID on their observation and another person corrects it, usually the observation will be moved there anyway.
I think it depends on the size of and the variability in the taxa. My filter is usually set to “Coccinellidae” because even just having it set to “Coleoptera” would mean that I cannot ID the vast majority of observations. The issue is even more pronounced with flies, where I can only ID a few species (Polyporivora spp. and a few other Platypezidae).
However, bird IDers and my local Bryophyte IDers seem to have their filter set to “Aves”/ “Bryophyta”.
Personally, if I want to ID plants, I set my filter to “Magnoliopsida” because I’m neither a taxon expert on a specific group and can also only ID most species (of those I can ID) to family and not to species or genus unless really taking my time and going through my keys more thoroughly.
It does in my browser (Safari). It says “Unknown” as the title and smaller underneath: “Placeholder: xyz”. I agree placeholders should be handled differently, though.
I find the CV a very useful tool as a starting point into research.
As it shows a list of similar species, I can start by looking at observations for every one of them and figuring out the taxonomy and the relatedness of the individual results. I identify for the largest monophyletic group that includes all the species I have deemed similar enough to my own observation and look at all the species in it observed in my area. In a separate window, I can try to find a key and then work through it with Research-Grade iNat observations used as references.
(I don’t do this for every observation, of course. It’s a lot of work. I just do this on the ones I’m really curious about.)
Alright. This is new information to me, which makes that point of mine moot.
That is an awesome outcome when it happens. Like I just learned there are temperate rainforests, and it was an euphoric experience that expanded my thinking. Like years ago when I learned that not all deserts are hot. I would not rob anyone of that joy. I just think it should be something that is in the filters either way; used or not used if it’s the only ID.
I think you misunderstood the point I was trying to make (or I misunderstand the point you are trying to make). The variation in methods used to come into a conclusion of an ID is a variable based on personal preferences, that makes the data quality inconsistent. I am arguing, that the user of the data has to understand and consider where such data can be used. There are no criterion for methods used in a project such as this.
And I am referring, that there seems to be a continuing trend of reasearchers and other professional users of this kind of data to complain on these forums, who seem, to me at least, misunderstand the constraints and limitations of a citizen science project produced data and how the system itself works. Might be I’m taking thread openers for discussions as complaining out of ignorance to the platform’s concept.
And this is an issue all of it’s own, that has been discussed at least since the 90’s. Do various social constructs and social environmental factors affect scientific work and focuses or not, and if yes, to what extent and how. The grants have people behind them, and they have their own interest groups to heed to. And even the peer-review system creates networks of people peer-reviewing the others in the network, and thus have a vested interest in the approval of others.
And yes, this is pretty much the gist of it. The science community isn’t as objective or detatched as many of their number want to make it seem, or even believe in it all good faith.
Saying I don’t understand the need, doesn’t mean I don’t use it. 
Actually, I use it pretty much identically to you, including having a separate tab or two for comparing and referencing. Where I start this process, though, is when I already have a rough idea of what I’m looking for. Otherwise I might go for something as wide as Angiospermae, if I it’s obviously one, or something like Asteraceae or Brassicaceae as as a rule of thumb, they have pretty distinct flowers. If I think it might be something, but am hesitant, II add the species to the comments and add I’m not sure but I think it’s likely this. This is what I mean by not seeing the need for it: you can use a coarser ID if not sure, but as this thread shows, it has its pros and cons.
The main thing that peeves me is, if the person using CV isn’t doing anything else, but clicking on the suggestion without a second thought. I was told in a comment to my comment, that CV is a project on its own, that uses the data in iNat to teach the recognition software. That made me understand its presence.
Reading through this topic reveals that the ‘problem’ is only applicable to plants. For animals it is in all cases useful to make broad IDs like spiders, insects, even Arachnida or Arthropods.
At least I do use these filters (or go to broad category projects like insects) to find things I am intrested in or enthusiastic.
I do myself such broad IDs for taxa I am unfamiliarwith when looking through observations filtered by places.
Re the plant discussion, I’d like to say, these suggested projects are probably unknown to many expert identifiers, so spending time with these may also be a waste of time.
Perhaps it would be most useful, to start a new topic and ask expert identifiers, how they filter for the taaxa of expertise.
I am afraid that iNat will break first. If identifiers flee in despair (too many poor quality photos, too many joke IDs from hit and runs, too much too many for that taxon specialist to keep up with)
then iNat will become another Facebook or Instagram or whatever pile of pretty or not so pretty pictures. Without IDs iNat isn’t iNat.
It is an interesting conflict between determined observers who see All the Plants, and identifiers who are looking at The Elephant. Since iNat has so many plant obs, we need to find a better way.
I’ll go start that new topic. But I’ll warn you that I will ask for data on prolific identifiers, not expert ones, because while I am prolific (247,000 IDs), I really do not consider myself an expert in anything.
but, you do realise that there are neglected taxa AMONG the plants?
This could be the second obs on iNat while I wait for my 2 taxon specialists to weigh in.
This whole thread is actually pretry discouraging. I think of myself as a very basic user, who really doesn’t know anything about biology beyond that’s a tree, spider, ect. I started doing this recently, mostly to id what plants were in my back yard and as sort of a landscaping aide. The citizen science part was just a bonus, maybe I’ll spot an invasive or rare species and that info would be useful to conservation people. What I’m taking from this thread though is that too many people misidentify stuff, even if they aren’t sure themselves. Several of you seem to know who on here actually knows what they’re doing, but there’s no way for the average user to know. It makes the suggested id’s unreliable, and sort of defeats the purpose. If we can’t trust the community id to be accurate, then what’s the point, for the backyard observer or the researcher?
I think you have to keep in mind that many times what gets discussed in the forum are the outlier cases, where everyone’s trying to tinker around the edges of a really well-functioning operation. My experience has been that the vast majority of IDs I receive on my own observations are correct, and I suspect that the vast majority of the IDs I make are correct as well (I screw up sometimes, but then, I’m human).
If you are fairly active on iNat, you’ll start to get an idea of who knows what for your area. You can also always look at someone’s profile, in case they mention that, say, they taught entomology at XYZ University for 37 years and published 42 papers on the taxonomy of Beetle Genus ABC. Or look at how many IDs they made. Or look at the leaderboards for the taxa you’re interested in in your area.
In other words, do not be at all discouraged because of all these discussions on the forum. iNaturalist identifications are not perfect, but then, neither are the identifications made by professional experts on plant specimens in herbaria.
As there are among any of the other kingdoms. I am sure these problems are a lot more pronounced where plant diversity is higher and perhaps less well described than in central Europe.
However, I can only speak from my experience and that is that of my 1444 plant observation 631 have reached RG (≈43.7%), whereas my 3088 animal observations have yielded 1151 that are RG (≈37.3%). Those percentages are pretty comparable, IMO. Which aligns with my feeling that I get about the same rates of IDs for plants and animals.
Then compare that to fungi and I have 645 observations of which 155 are RG (≈24.0%).
My fear is that if we just leave all unidentified vascular plants in „unknown“, that will make it harder and more frustrating for the few identifiers to find a lot of the even more obscure taxa (lichens, rusts, mildews, algae, etc.).
I think the core of the problem is lack of plant education. Many people know „that’s a plant“, but wouldn’t be able to tell you „that’s a member of Fabaceae/Brassicaceae/Lamiaceae/…“, whereas most are able to differentiate a dragonfly from a butterfly from a beetle. This isn‘t an iNat specific problem, so I don’t think we should expect iNat(ters) to provide the solution.
I believe for now the best (and fairest) answer to this problem is trying to teach as many potential identifiers to ID the most common and easy to ID plant families. Perhaps writing a copypasta with a link to a relatively simple to use resource that plant IDers could link to under every ID could help in that regard?
If there is any interest in this, I‘d be willing to try making one. I wanted to practice my botanical illustration anyway. :)
The forum is where the hardcore iNat addicts gather! Welcome.
Spider works well.
Trees, ah, that gets complicated. Pine trees don’t even flower or fruit. We all know what a tree is, but taxonomy does not ‘tree’.
iNat and biodiversity is a learning curve, ultimately long and slow, but with hiccups and derailments along the way. The very best identifiers are the kind ones who take time to say … this is … because. But first you have to prove yourself interested and willing to learn from them. Ask questions, then use what you learn to help ID the backlog.
I promise. One day you, will be the one to say, wait, that is not a …
There is a lovely German expression - aller Anfang ist schwer - It IS hard at the beginning (clunky in English)
Don’t be too discouraged. On average, Research Grade IDs are 95% accurate, which is pretty high. All information has limitations and can rarely be trusted 100%. That applies to museum collections, herbaria, scientific databases and any other large collection of data you care to think about. The nice thing about iNaturalist is that it is relatively easy to find and correct errors, unlike many other databases.
@DianaStuder What does this mean? If you can answer my question, please do.