A relevant passage or two doesn’t make a relevant article. Imo, neither of the articles is very clear about its intended reader. There should be a separate series of articles for observers and identifiers. The initial article for new observers should be very simple and straightforward. It should discuss appropriate guessing using the algorithm and probably not much else.
Where is the Similar tab? I searched and cannot find that. Sounds like something I would really like to use.
The most correct answer is obviously: yes.
The best way to use iNaturalist is to use all three resources! With practice you will learn that each has its place, each has its limitations, and each has skill involved in using it most effectively.
The CV is really helpful as long as you use it as a tool and not as a magic trick. With practice, you will get a sense of when you can cautiously rely on it and when you really, really can’t.
Field guides and keys and other resources are great, but the taxonomy doesn’t always exactly match, and the field marks listed are not always adequate for our purposes, because identifying from photos is not the same as identifying in the field.
Community experience helps to bring it all together. With practice, you will start to learn who to ask for help, and how to make it easier for other people to help you. One very simple tip is to crop your photos appropriately, so people can immediately tell what you are trying to identify. (Searching through five hundred leaves for the bird is pain. You’ve already done it, so why leave it for identifiers to do again?)
To use iNat’s CV – click Compare – then choose Visually Similar.
Or on the relevant taxon page – click Similar Species
Once you have a tentative id for your observation, it’s a good idea to look at the distribution map for that species. If your observation is from the coast of Maine and all of the iNat records are centered around Japan, you are probably on the wrong track.
Lots of great suggestions here. A couple more suggestions for when someone does disagree with your ID:
- Don’t feel bad, everyone is learning!
- If you think they are probably correct but you don’t have the expertise to say for sure, please withdraw your ID. That’s generally the best practice rather than agreeing with the new ID or not doing anything. If the new ID is NOT correct and you agree with it, then it will take 6 further ID’s to resolve it, which is a frustrating situation. If the new ID is correct and you leave your disagreeing ID up, then it will show up as a higher level taxon and take longer for identifiers to find and fix. You can always return to it (and agree or disagree) after you’ve had time to learn some more.
Have fun!
This has always (8 yrs on iNat) been my usual procedure: (1) Take iNat’s AI suggestion when I’m in the field. Getting a genus or family completely wrong seems to provoke a faster expert ID if it’s got a species name than if it’s left at genus or family. (2) When I get some time on a desktop machine, return to that observation. Look at iNat’s photo library and descriptions of that taxon. Do you agree that it’s a good fit? Maybe you want to make a change. Also check the distribution maps for similar species. Do its geographic location and habitat make sense in this context? At this point you will have learned about at least one new species, even if the ID turns out to be in error. (3) If someone tells you it’s a different species and you still don’t understand why, then ask them. Almost everyone who does ID enjoys getting such questions!
If you are taking the route of buying books, avoid books with unrealistic titles such as The Insects of North America or The Complete Guide to British Wildlife. They are likely to contain pictures which look like the specimen you want to identify, but they won’t mention that there are a dozen species that look the same until you get down to microscopic characters.
Welcome! I use all of the above. I rely quite a lot on field guides, but also try to take classes whenever I can. Taking a field course on lichens really helped jump-start my identification skills with those organisms, and now I see them everywhere.
I have been impressed with the new iNat app computer vision too. It has been remarkably good at identifying uncommon plants in remote areas, juvenile lizards, birds if I can get a good clear close photo, and higher insect groups, for example. I have years of a base to work from to double-check the CV, so on the few occasions it gets things wrong, I can disagree with it. But it has greatly improved recently.
I have also become much more comfortable with choosing the higher taxa that the CV suggests, knowing I may not have the skills to identify certain groups to species. I was happy to choose even a “subtribe” of plant hoppers, checking on the internet and my field guides that this taxonomic level narrows it down enough to get the ID in the ballpark. It has been fun learning the higher taxonomic groups better. If you need more on the basics of taxonomy, check out https://www.britannica.com/science/taxonomy/Nomenclature
Have fun!
I generally let CV point me in a general direction and then look at its suggestions and the Similar Species of those suggestions. Sometimes, CV is hilariously off-base and I’m on my own. If I don’t feel confident about any of its suggestions, I look upward in the taxonomic tree of each to see if all of the suggestions meet somewhere, like a family of moths or an suborder of insects.
Keep in mind that CV only knows the taxa that it has been trained on. The program only trains on a species when it has a certain large number of Research Grade observations. That means it won’t suggest things that are uncommon or infrequently observed in your area. I sometimes figure out a high level taxa and then look at photos of everything in that taxa in my area. Do that by clicking “View More” on the taxon page and then selecting “Taxonomic” in the Grouping dropdown list at the top. That has the added benefit of showing you things you never knew existed but, which you may be more likely to see in the field if you’e seen photo of them before.
If you have good, crisp photos of insects or insect-adjacent critters (spiders, etc) in the USA, you can also use the Bug Guide ID web site at https://www.bugguide.net. (Bad photos end up in the “frass” pile.) I use BugGuide for things I’m in a special hurry to figure out or for observations that have stalled in iNat. Some BugGuide identifiers also identify in iNat. I point the BugGuide and iNat observations to each other in a comment.
Well, maybe if you’re in England where the flora and fauna are pretty well known. South America, not so much. I’m pleasantly surprised to see a species with enough observations to make it into the CV pool. Very few tropical Eastern Pacific mollusks do. Many of them have fewer than ten observations, not because they’re terribly rare, but just because there are few observers. With luck, the model will suggest the right family or even genus, but often it is way, way off.
Lots of good advice here, and I moved an off-topic conversation into its own thread. I’ll note that we just released Identification Tips, which is a way for the community to search for, nominate, and vote on helpful ID remarks. It will probably take some time to populate, but I think could be a good way to find information that will help you with identifications.
To confirm your suspicions and then get back on track since this thread got split - The mobile apps kinda suck (Well… they really really suck) at allowing people to make IDs for others. I think that is a massive source of the difference between amount of IDs and of observations.
I recently made a Reddit thread (And it got pinned in the inat subreddit, yay) for people that want to make IDs but are scared of doing it wrong (and it gives some tips for OOPs’ question as well) and I mention to use a PC browser for this reason - Honestly, even when I really tried to do IDs for the period of time I was without a PC, it was incredibly frustrating.
From your description it seems the iOS app is very similar to Seek and forces you to ID - I was not aware of that. In Android you upload/take the picture and then you fill each detail. When you go to fill the ID it takes you to the CV and you can write the name of the taxa or pick a suggested species, but if the person doesn’t click the ID field at all and just submits the picture, it uploads as unknown. In Latin America iphones are relatively rare (So everyone is on the Android app) so there are oogles of observations uploaded as Unknown. It’s interesting to know how much of a difference it makes in other countries.
I’m sorry that you moved my initial post in that discussion because I think the point I was trying to make is relevant to this thread – namely, that it is desirable (from an IDer standpoint) if observers spend a little bit of time thinking about their IDs rather than just picking a CV suggestion fairly uncritically and hoping to be corrected if it’s wrong.
This doesn’t have to mean doing extensive research, but looking at the suggestions and assessing whether they seem plausible, as well as a quick check of observations at the next-broader taxonomic level can often provide a sense of whether there are similar species and whether there might be identification challenges that aren’t obvious from the suggestions. e.g.: If a considerable number of observations are not getting confirmation, this is often a clue that it is difficult; if several of the suggestions or the suggested species and other related species look very similar to you, it is a good idea to be more conservative with your ID.
In the field I find that the app is less confident than at home on a computer. Some of it has to do with signal strength and some to do with editing. If I wait to crop I can also chose the pics with the best focus and put them in the best order (whole organism and clearest first…). Several views of bugs (front, back, side, top) and plants of all sizes really helps increase confidence (and helps with RG ID). Also, I’m happy with my confirmed (RG) species count a little less than half of observations, which changes daily.
Good advice, thank you.
Learn to distinguish basic insect orders/other groups of your interest. That isa good start.
Keeping this topic alive because I’m writing a massive reply.