As an identifier, I get worried sometimes that we’re going to lose some of our experts because they get frustrated about the platform or the activity on it, but I think it’s more important that we help people figure out how to use the site/app and that they’re enjoying and learning about nature while they do it. It seems to me that a lot of people (myself included) learn more from corrections than from simply using an identifying tool. What an opportunity for some of us to share a little knowledge and build the community! Accuracy IS of utmost importance, but there’s no reason we can’t be respectful and have a great time while we’re working on it. And the fun goes both ways—I really enjoy your observations, @walkingstick2!
8 Likes
Well said walkingstick2, I hope you keep posting. I am like you, and if we tried to learn the answers to all our observations by researching alone we would not have time to contribute much observation. Like you, i try to learn as much as I can from others’ IDs and the information they provide so that I can gradually ID better myself, but in the meantime all my observations are useful to me as a record of the present ecology of a place (evn if its just trees versus grass, for example), and for comparison with other obs of the same plants. Some of my obs are useful to others now and again.
Over time people with more knowledge, including my future self, come along and improve low level IDs. I try not to confirm unless I am very sure, no matter how long it takes.
3 Likes
Yes this is true, and I second @sgene’s advice. Best to put in the finest-level ID that you feel reasonably comfortable with. And if that is “fungus” or “lizard” then that is still much better than nothing at all, or than an obviously wrong suggestion from the app.
The “Computer Vision” suggestions are kind of like spell-checker suggestions – they are often right, and sometimes spectacularly wrong. Our day-by-day observations help it improve slowly but surely.
For more about IDs you might find this discussion of interest (and so we don’t get too off-topic on this one). Glad you have joined us on the forums!
7 Likes
This is incorrect! “Research Grade” may be somewhat of a misnomer, but it is clearly defined as an observation for which at least two user (accounts) agree on the ID, and for which there is no significant disagreement to counter that. Hence there is no way that an RG obs can fail to “actually” be research grade in the defined sense.
You’re not really referencing the connotation of @nathantaylor comment. Yes, an observation can reach iNaturalist’s definition of “Research Grade,” but that doesn’t mean it is actually a correct identification, which is what I think was the intent. I know this has been covered in multiple other threads before, so if you are interested in debating the terminology, it might be better to address it there than hijacking this one.
4 Likes
like this one, where the title of the thread involves Research Grade: https://forum.inaturalist.org/t/rename-research-grade-discussion-and-polls/590
1 Like
Thank you for your question! I have been wondering about reasons why people might choose the top computer vision suggestion. I think @sgene and @jdmore answered your question well. Do put in an ID, but the computer vision suggestions can often be wrong so it may be wiser to put something that you know for sure it is (it’s definitely a plant or a bird, for example).
4 Likes
Yeah, I’m always a little nervous when my observations go to Research Grade after just one other person has agreed with my tentative ID and it’s someone else who seems to be a newer, non-expert user. I guess I could start overriding it and checking the “this ID could use improvement” box. But when someone who’s a professional or a very experienced hobbyist confirms my ID I feel like I’m that much closer to being able to confidently identify that particular organism, while when it seems to be just another person going “yeah, that looks like the picture from the algorithm” it’s harder to know if there might be three other organisms in the area that also look like that picture.
9 Likes
Thank you very much for this message! I couldn’t agree more. There are tens of thousands of regular users of iNaturalist, and in my mind, it’s ok that different folks have different ways that they want to use this tool. I regularly need to get myself out of my bubble of “this is what I want to see, and this is how this tool should be used!” and recognize that different people may use this tool differently. And that’s ok! As with most things in life, it’s probably good that we take a step back every now and then and look at the big picture.
Example: a dear friend of mine takes pictures of cardinals – and nearly every cardinal she sees every day. The scientist in me thinks, “Whew, this is redundant…” but the fellow naturalist in me realizes, “Wow, she must really enjoy cardinals! Cardinals are pretty cool.”
Other folks document different individuals of invasive species in a small area. Again, I may think it’s redundant, but they’re using the tool in their own way. That’s fine, yeah?!?
I also worry a little about the “expert fatigue” from some folks that get hammered with lots of ID requests or are discouraged with poor photos or incorrect angles or whatever the case may be. I try to reach out to these experts with messages of encouragement and gratitude, and I hope that the observers do this as well.
I lead quite a few public trainings on “how to use iNaturalist” and I get this question a lot: What’s the right way to use iNaturalist? My answer is always, “If it’s helping you enjoy nature and engage with fellow naturalists, you’re using it correctly.” :)
Just my two cents.
Now, I’ve go to go take some cardinal observations. :)
25 Likes
My point was that there should be no assumption that an RG observation is correctly identified. The only assumption should be that RG observations are more likely to be correctly identified than those which are not RG. It is simply no more or less than a level of confidence above non-RG. Nobody here has yet demonstrated that the problematic cases are frequent enough to undermine the step up in confidence.
Now I see why you seemed to misunderstand my point. It is not about whether they are RG or not, it is about whether they are dropping out of the needs ID pool (default Identify parameters) and therefore getting less of a chance to be corrected. And yes, another way to solve the problem would be to open up those parameters and include RG obs, but then the needs ID pool would be huge!
1 Like
That may be your point Mark, but I was replying to someone else who was making a different point.
As for your point, I personally try to pick taxa that I know and look quickly thru all obs identified as that taxon to check if everything seems to be in order, regardless of RG or not. I see your point, but it requires a trade off, as there are significant disadvantages whatever you do. At the end of the day, the best approach seems to me to just try to improve data quality over time and we have to live with a low level of misidentifications (which is not a problem restricted by any means to iNat and the like, but every collection in the world and every publication!)
2 Likes
Point well worth remembering here! Collection curators spend entire careers not finding them all.
2 Likes
I have even found people sharing their ids for observations a couple of years old - and its a real pleasure to be reminded of these puzzles and sometimes resolve them. I only confirm an id where I can also judge that it’s correct based on my own limited knowledge / time spent peering at descriptions online
4 Likes
I’d like to hear from researchers about how they take a data set from iNat and what they do about the overly redundant cardinals and the indistinct photos in order to make it useful but not introduce bias when filtering.
Personally I found this place really hard to follow - In most sites a forum page is somewhere where one posts questions to get answers so that is where I posted a picture. I was rapidly shot down for that which was annoying as was finding my way around. I’m just an average outdoor person one that doesn’t really give a toss about latin names ( even though I did study Latin and Biologyat school) and prefer to use the Maori or common names but my main reason is I want to identify a plant so I can find information about where to replant it mostly : sun/aspect/soils /drainage/wind etc and that all starts with an I.D.
1 Like
Hi Simon,
I take it you are referring to this?
https://forum.inaturalist.org/t/plant-identification-please/2439
you did ask a question, and it was answered, and I feel it was in no way “shooting you down”.
The first response pointed you in the right direction, and even followed up by asking you how you got to the forum to post your plant ID request instead of iNaturalist itself, which I can only assume to be an effort to evaluate whether something is misdirecting people to the forum for those type of enquiries.
The second response you got was to point out what the forum is actually for, as specified in the welcome message you would have got when you joined the forum. Again, a polite response and hardly “shooting you down”
then there was my response, which was to acknowledge that it was a simple mistake, and that to get your ID a little quicker you could start the ID process off by going as far as you can tell yourself, eg plantae, so the specialists for that area would get to see it quicker. Again, I thought I was polite and not “shooting you down”.
In my book, that makes the response you got pretty thorough. In fact, I feel at every stage of that you were encouraged to “keep flying”, so I have no idea where you get this notion that you were being “shot down”, and I think you may have your expectations set a little too high :)
There are some tutorial videos on the help pages, they might be a better place to start.
This might be reading between the lines too much… but “Aoraki-climber” and “where to replant it”… you aren’t taking plants from a National Park, are you? If so, then you definitely need to be shot down!
4 Likes
Correction: the definition must include: that it has to be “wild” and that planted or captive observations cannot be either “Needs ID” or “Research Grade”. No matter how many agreements there are to an identification.
1 Like
As one on both ends. Although I submit thousands of observations to iNaturalist, my primary interest is using the data. These are some of my uses.
** To update distribution maps: as a rule we know a lot about charismatic rare species. But it is the most common species that we often know least about. They are not “important” and take up space in museum/herbarium collections. Furthermore, knowledgeable people ignore them as noise. But beginners often provide valuable new distribution data. And ditto people visiting new areas who are unfamiliar with the species and want a quick ID: these data are invaluable. These data are useful for modelling all sorts of things, from Red List status, to how species will “move” with climate change, and how many populations are in protected areas.And also for Environmental Impact Assessments - both doing them and checking them (when a survey says an area is devoid of threatened species, and the iNat data says that there are lots there …).
** To review Red Data Status. In the Cape Flora we have three quarters of southern Africas threatened species (in < 5% of the area). Monitoring these is impossible without citizen scientists… We have trained volunteers (CREW) to go out and check up on thousands of populations in Cape Town regularly. These not only post pictures, but also population counts, alien threats and other issues regarding these populations. This allows our Red List status for species to be kept up to date, and not be 10 years old and thus irrelevant in a rapidly urbanziing city. Also many people observe threatened species without realizing it, and our CREW volunteers then request them to provide the data we need, even though they often (and understandably) cannot remember the population sizes and extent.
** Every August thousands of Endangered Western Leopard Toads have an orgy over two or three days, usually in rain, and usually during evening rush hour traffic., Hundreds of volunteers rescue these toads from the roads. The iNat smartphone app is useful for them to quickly photograph and sex the toads, with the proviso that rescuers need to ensure that they dont end up flat on the roads as well. We can use the data to individually ID each toad and calculate all sorts of statistics: longevity, breeding site fidelity garden fidelity (the toads live in gardens for 50 weeks of the year), and local population size.
** Surveys. Keen and interested citizen scientists can join special projects and contribute specific data to them. There are lots of examples, but for instance - to assess the potential impact of fracking in the karoo, anyone - including farmers and tourists - could add data to this project - either opportunistically or in a structured way that allowed calculation of density and abundance. Lots of people did.
** Detecting new alien invasives and the spread of pests. Although many citizen science may not realize it they collect useful data on ruderal and invasive species. These include first instances and rates of spread from casual data. These are most useful in working out how to curb (when possible) a threatening species, and how much it will cost to control them. This includes the spread and success of biocontrol measures. Although most users are initially unaware of it, most biocontrols are visible or their effects are) in photographs of aliens if they are present.
The important thing is to get away from the ridiculous notion of photograph quality. Some species are easily identifiable from the most blurred pictures. Others cannot be identified below tribe even with dozens of exquisite and technically perfect pictures. The beauty of the interaction between specialists and beginners is that interested beginners can rapidly be informed (indeed “trained”) as to what features are needed in photographs to make identifications. Not only does this benefit those making the IDs but it makes the group more interesting to more curious beginners and starts them on the road to becoming proficient at identifications themselves.
**
Some projects illustrating these, including some fun projects to join (if you are in southern Africa- look at your area for your local projects):
https://www.inaturalist.org/projects/alien-early-detection-rapid-response-s-afr
https://www.inaturalist.org/projects/ant-atlas-of-cape-town
https://www.inaturalist.org/projects/biogaps-s-afr
https://www.inaturalist.org/projects/tree-pests-diseases
https://www.inaturalist.org/projects/vegmaphoto-s-afr
https://www.inaturalist.org/projects/illegal-harvesting-s-afr
https://www.inaturalist.org/projects/strandings-s-afr
https://www.inaturalist.org/projects/blue-community-s-afr
https://www.inaturalist.org/projects/wlt-monitoring (2000 observations still to be added - data migration issues)
https://www.inaturalist.org/projects/spoor-signs-s-afr
https://www.inaturalist.org/projects/used-medicinally
https://www.inaturalist.org/projects/crew-site-sheet-s-afr
https://www.inaturalist.org/projects/redlist-s-afr
https://www.inaturalist.org/projects/roadkill-s-afr
https://www.inaturalist.org/projects/harlequin-ladybeetles-in-southern-africa (this project contains the first records of this species in Swaziland and Botswana)
https://www.inaturalist.org/projects/dying-fynbos-s-afr
https://www.inaturalist.org/projects/reared-from-larval-stage-s-afr
13 Likes
Distribution maps, combined with other records, via www.gbif.org. Duplicate records don’t matter as not trying to extrapolate abundance. Indistinct photos don’t matter as unlikely to reach Research Grade. Pot plants that haven’t been tagged as “not wild” are a pain.
2 Likes