I know the general topic around this question has been discussed multiple times before, but I’m at a loss regarding this situation. I will try to avoid details allowing to identify the user harboring this behavior but I can edit the post if it’s too specific nonetheless.
There is one very prolific user (we’re speaking in a scale above 30k observations) uploading a gigantic amount of duplicates. Sometimes, the same organism or group of specimens can be duplicated up to 10 times. The duplicates are not identical pictures but multiple angles of the same organism. Most, if not all observations are identified with CV and often completely wrong or unidentifiable with the details given. In a few cases, there is even no evidence of organism.
I tried reaching out by commenting on some of the observations asking if they could merge the duplicates into a single observation but I never got any answer.
The problems caused by duplicate observations have already been discussed thoroughly, but this case is especially troublesome as most of the observations are for groups massively lacking identifiers and adds a lot of work to correct the identifications and annotations. Less importantly, it also impacts the observation leaderboards given the scale of the duplication.
all I can say is stringers like these have been prominent throughout birding (don’t get me started on swallowgate) and really all sub groups of naturalists pretty much as long as data collection or platforms like inat and ebird have existed. the implications imo aren’t that big on something like inat, bc lets be honest who gives a shit abt the leaderboards the more impact is the data collection since it swarms and bloats the database when it shouldn’t, were all here (or more should be) for other, much better and honest reasons. wether this user is unbeknownst to this or does know, thats really up to the moderators to decide to take action but until then, this sounds corny but sit back and relax lol👌
Thanks for the answer. However, the leaderboards are a lesser issue as I mentioned in my original message. The problem is mostly the amount of work properly identifying and annotating the observations for identifiers. Typing 10 times the correct name takes obviously 10 times as much time for the identifiers than typing it once :)
From previous posts in this forum, duplicate observations are discourage, but not actually forbidden. If the ID’s are correct, just confirm them and move on. Creating duplicate observations for each individual in a flock of birds, for example, is allowed as well, and could potentially be useful for annotating different ages and sexes.
The problem comes when most of all of the individuals in the flock of birds are unidentifiable past family or even order level, and there’s 117 duplicate observations of blurry dots that may or may not be European Starlings. In that case, I recommend adding an ID as accurately as possible (maybe just “perching birds”) and clicking the “no it can’t be improved” option in the DQA (as long as you’re 100% confident that they can’t possibly be ID’d further based on the evidence shown).
I suggest creating a flag if you think it is extreme enough. Curators can determine if this particular case violates any guidelines and if there is a need to do anything. The flag could result in action, or be resolved with no action.
While duplicates are generally allowed, if they are happening as frequently as you describe. It could possibly turn into something actionable.
In extreme cases, certain things can become guideline violations for various reasons because they are extreme cases. As with many things on the guidelines, scale matters. Uploading 1 human observation is fine, uploading 200 is not and could result in a suspension if continued after a warning.
But it all really depends on the details. I know of a few extreme cases of duplicates where action was taken. But those are uncommon extreme cases.
(EDIT: I should mention that I came across the user and I looked at a few of the observations that Nick is making reference to. The activity does seem bizarre as many folks have tried very kindly to correct the behavior(s). Doesn’t seem like the message is ‘hitting home’.)
Interesting question. It doesn’t take too much time to learn how to combine several photos of the same subject (flora, fauna, or fungus?) into a single observation. It is easier and more efficient to combine the observations when uploading instead of after posting. This morning and afternoon I posted several photos as one observation, and two separate photos (of a Monarch and a Milkweed vine) as two separate observations. I did a separate post of two photos of what appears to be the same species of vine from the previous post, not to increase my post counts, but because the flowers were in different stages of blooming and the leaves, though similar, suggested a different species from prior experience.
As an amateur, I use iNaturalist as a learning tool, as well as to (hopefully) contribute to the quality of the database.
On a related topic, which might help to address some of the issues raised previously:
Is there a written or video tutorial that addresses How To create a simple, single photo observation?
… to upload multiple photos as a single observation?
… to duplicate an observation observation with multiple subjects into and revise the Identification of the species (i.e. split one photo observation into two separate species observations)?
… Add photos to an observation to provide additional views or details of the same object at the same place and time (e.g., a full plant photo, and then adding details such as leaf shapes and arrangements, buds, blooms, and if present at the same time, fruits and seeds, etc.)?
In other words, bending the twig in the right direction would take less effort than trying to prune a tree which has grown without constraints and restraints.
(BTW, I have seen some guides to identification within iNaturalist which have been useful. However, I’m not sure I would know where to look for them. Both technical training in using the database and specialized guides for identifying such things as different Milkweeds and differentiating between likely candidates, located where they are obvious and easy to find, could possibly relieve the burden of identification noted above.
On the Frequently Used Responses page, there’s a section for Multiple Observations for One Organism which links to a tutorial page, and that has a link to a video tutorial.
From what’s been said above, it sounds like this has been pointed out to the observer under discussion, without effect.
We do need a DQA for - Duplicate please delete - which would push the obs to Casual. Similar to NOT a Single Subject.
To the original question, my workaround is to filter out the offending batch, Mark as Reviewed, and move on. But this situation sounds like a flagable offence.
If it happens occasionally(I definitely upload duplicates on occasion, usually when I mistakenly think that there is a pathogen/fungus on a plant or forget that I took multiple photos of the same plant.); it’s fine. Even uploading 1000 photos of the same species as long as they are separate individuals is also fine, population censusing is incredibly important(although a common species like white clover would just be boosting stats).
If the intent is malicious with deliberately false IDs then that is suspendable. The AI is much harder not to use on the new IPhone app now, and I need to use it when I am in a location I’ve never Inatted before.
FWIW they did eventually give me a couple of thank you comments after I corrected hundreds of their IDs, so they’re not completely closed to communication maybe it’s an English issue, someone should maybe DM them in their language with step by step instructions on how to move photos of the same organism to a single observation when uploading, before anything more drastic is done.
There is a difference between duplicate observations and multiple observations for a single organism–and both are a problem. I think duplicate observations are much easier to handle and seem less frequent. The same picture is posted again, usually only a second time; in those cases, I copy the URL of one and add it to a note on the other, saying something like “This is a duplicate of [URL]. Please delete.” Sometimes only one has been IDed, so I try to be sure I mark the other one for deletion.
The bigger problem that’s being discussed here is when, for example, someone posts 10 pictures of a single bird in action as separate observations. I’m no expert but I’m sure that is skewing the data. No, ten of this rare bird have not been sighted! The Frequently Used Responses page is careful to note that the responses for this case are intended for new users. That’s easy enough. But when the user has many thousands of observations, they’ve been at this a long time.
I have faced a few similar situations to this (but with no language barriers). The main issue for me is not so much the duplication, as the continual lack of any response. At some point, this may start to run contrary to the community guidelines regarding “good form”. Cases like this seem quite similar in spirit (but obviously not in every detail) to this guideline:
Add accurate content and take community feedback into account. Any account that adds content we believe decreases the accuracy of iNaturalist data may be suspended, particularly if that account behaves like a machine, e.g. adds a lot of content very quickly and does not respond to comments and messages.
The guidelines are now more tolerant regarding duplication, which I suppose we all have to accept (up to a point, anyway). However, never responding to polite requests without providing any good reason, seems an entirely different matter. Of course, it may be that the user is simply unaware of how the notification system works and how to use it effectively. But either way, it seems legimate to raise a flag for cases like this, and require the user to make some kind of response. That should at least allow other users to make a properly informed decision about how to treat the user’s observations in future. If the eventual outcome is to just let all the duplicates pass, so be it - at least we’ll know we’re no longer dealing with a machine.
to avoid becoming that duplicate obs guy, for example if I record a bunch of sounds of a robin in different places on different days is this OK? I have no means of knowing if this is one robin in different places.
There seems to be an implication duplicate = different takes of same individual at the same time, but if multiple takes of a common species in different places is deprecated then that really should be made clearer somewhere ;)
This is not a problem, as iNaturalist data, in general, are not suitable for assessing abundance. The issue here is the additional burden on identifiers (and perhaps also the environmental impact of storing large numbers of redundant photos when one or two would be enough to record the presence of that individual on that day at that location).
Your example would not fall under duplicate observation, even though it’s the same species. The issue pertains to the exact same time, date and exact observation being posted multiple times.
maybe it’s an English issue, someone should maybe DM them in their language with step by step instructions on how to move photos of the same organism to a single observation when uploading, before anything more drastic is done.