I agree with @ocean_beach_goth. I always try to identify at least as many observations as I submit, but mostly focus on welcoming new people in my state and identifying unknowns as plants, insects, or other top-level categories out of fear of not knowing all the possible local species. I’m an experienced gardener and have taken a class on nomenclature. If I knew what resources I could use to usefully identify native wildflowers and trees in my state, I’d spend a few hours a week doing this.
Like you I want to learn plants native to my region. I see from your profile that your state is North Carolina. Mine is South Carolina. You might find this project a good starting point since the Carolinas have some plants in common. https://www.inaturalist.org/projects/south-carolina-native-plants
Took a long time, but I got through all the San Francisco Bay Area Asteraceae. Moved on to the Puget Sound Country Asteraceae. I am developing a hypothesis that every region has that one non-daisy that people think is a daisy. So many California Buttercup identified as “Asteraceae.” That taxon does not occur in the Puget Sound Country, so what do I see there? Common St. John’s-Wort identified as “Asteraceae.”
People put too much trust in the apps. Computer vision works well for certain organisms, but not well enough for others (like plants).
The pile will only get bigger and bigger. And besides that; there is also the not so small matter of the many hundreds of thousands of research grade obs with the wrong ID.
The pile does indeed get bigger and bigger. I was discouraged by that for a while, but I’ve tried to reframe my thoughts to realize that the more people use iNat and the more observations that are made are indications that this platform is a success. A runaway success, in my judgment. The more people use iNat, the larger those piles are going to grow.
If we want people to understand and appreciate the biological complexity on earth, iNat is a great tool towards that goal. Those of us who make identifications are providing positive feedback, not only to the biologists and naturalists who already know what they looking at, but also to people who are just starting to learn about the natural world. Hell, it’s only through iNat that I’ve learned - in the past month or so! - that Slime Molds are not Fungi. And I was a professional biologist before I retired.
Being overwhelmed and discouraged at the never-ending pile of Needs ID observations is completely reasonable. But are we accomplishing a goal that’s larger than just completing the task of “finishing” the identifications? I think we are.
The more data we have, the more data in need of ID we will have and the more data quality issues we will have. When i started using iNat in 2011 i would literally review every plant observation that anyone added. Which was barely any. It was awesome being on the top of the leaderboards but kind of an echo chamber. This is better/ The data is never going to be perfect, by any means, and will always require significant review if you’re using it in a scientific publication. Remember this is a group field notebook, not a heavily curated herbarium. And even still, heavily curated herbariums have plenty of errors too and often IDs aren’t ever verified by multiple people at all.
Yeah, I’ve seen a few herbarium specimens that were mis-identified, even though they were collected and IDed by very competent professional botanists. And I haven’t had a reason to look at lots of herbarium specimens, either. So, yeah, even experts make mistakes, much less people like me. It is really great to have people looking at and correcting my IDs on my observations. I’m hoping to learn more plants in 2022, so I can repay the ID favors.
If the NeedsID pile begins to consistently get smaller, it will eventually disappear. iNat would then be a very different experience for identifiers. You would have to wait for things to come in to identify, and then other people might well get there before you while you’re at work (or whatever). It would become a much less dynamic environment, with fewer learning opportunities. I’m not saying the current non-equilibrium between observations and identifications is ideal, but let’s just remember that the NeedsID Pile is what keeps iNat ticking. There is always a mountain of sweeties to get our teeth stuck into. Viva la NeedsID Pile!
Agreed about physical collections such as herbariums. Unless an expert in a taxon visits the collection or requests a loan of material to examine, some specimens can sit untouched in a cabinet and remain incorrectly IDed (or not IDed) for years or decades. Although digital photography and scanning has made access to images of these specimens somewhat better.
If there’ll be no needs ID as of now, people upload so much, it won’t take long to fill the pile in, some groups are just not checked and will grow no matter what until someone starts reviewing them. I think more free time would mean more opportunities to check old ided stuff and find mistakes.
As observer I really can’t care too much about iders need’ to learn by those observations if that means I will have them stuck at needs id forever.
My thought experiment was kind of hypothetical - the discussion above just made me try to imagine what iNat would be like without a NeedsID pile. I suppose the counter to that is that if we ever ran out of things to ID we would just need to go out and do more observing!
Like it is for birds? Just kidding. I’ve actually never looked at Needs ID for birds, but with the way my personal bird observations are treated, I imagine the pile is relatively small.
Ha! Yes I would imagine the same. Perhaps surprisingly there are 663,000 observations in the Aves NeedsID pile! But that is only about 5.6% of the total observations. Compare with insects - twice as many observations, but nearly 50% in the NeedsID pile (just over 10m!). Perhaps the birders have found that perfect equilibrium?
Ah man I’ve got what I think is a merlin that’s been languishing in Needs ID for months :P
Sometimes pushing the ID to … I think it is this, so long as you follow your notifications, works.
You either get a prompt - Not a merlin, or agreement.
There are 71,000 stuck in class Aves, compared to 11,857,019 overall, so more like 0.6% are actually not IDd beyond ‘bird’. It seems to me that if an observation in class Aves isn’t IDed within an hour it is either genuinely hard (like tracks, scat, bones, eggs, or a nest) or there is something wrong with the observation, or both. And a lot of that stuff will only get out of ‘needs ID’ when someone checks '‘cannot be improved’.
By comparison 1,870,816/35,857,963=5.2% of ‘Plant’ observations are at class or higher. Honestly, that seems pretty good given how many are just a leaf or landscape shot, but it clearly isn’t ever getting to the level of ‘Aves’
Well, now that I’m working on Asterceae for Vancouver Island, I’m not sure the computer vision is even the problem. Adenocaulon bicolor vs. Mycelis muralis seems to be the V.I. equivalent of the Bay Area’s Achyrachaena vs Uropappus – lots of people wishing that their Mycelis observation was Adenocaulon. The thing is, the default picture of Adenocaulon bicolor shows the observer’s hand turning over a leaf to display the white underside, which Mycelis muralis lacks. If people would just do that simple thing, they wouldn’t make that identification mistake.
There are so many plant observations where I wish the observer had added just one more photo, showing, for example, the number of needles in a cluster of pine needles, or a nice clean leaf from that oak tree where they took a photo of the bark, or (conversely) the bark of the tree where they took a photo of the leaf of what might be a Hop Hornbeam. And so on. Short of writing comments on every such observation, I don’t know how to get this sort of information out to observers.
I know I certainly appreciate that kind of feedback! Though sometimes it isn’t easy to get all the photos that would be helpful, like if all the leaves are out of reach. For some species, user commented tips are better than any available printed key.
In theory the site could be fine-tuned in some ways to prevent the pile from growing as much and help identifiers manage the influx better. I’m less concerned about CV than the general influx of new observations/IDs, since CV is more controllable by iNat in principle if it to cause significant problems. I’d also suggest iNat take steps like incentivizing photo clarity and cropping. Also, ID backlogs for specific groups can be posted on here to ask for identifier attention.