Platform(s), website, used from an IP address in Europe
URLs not relevant
Description of need:
In the last few weeks the preparation for observations upload has been hampered by some sort of refusal to get AI support for identification of species after approx 50 items in the batch. For observations of insects, a daily upload can easily surpass this number. To still upload the entire lot, I am used to just upload it all, anyway, even without ID. This will then go to the “unknown” category. I intend to resolve the identification issue after upload, one-by-one, because that is fast and works fine. However, in the meantime I get all sorts of “help” by others, which add generic taxa. This is unnecessary work for them, not really desirable to me and they would better do this for the really neglected “unknowns”.
Since also the search for my own “unknowns” takes time, It nowadays takes 10-20 minutes to come up with the correct updated set, I would really appreciate if my recently uploaded “unknowns” would be blocked for everyone else for a an hour or two. During that time, I would still edit and update entries in my observations.
Feature request details:
Hide any of a user’s recently uplaoded observation items in the “unknown” category for 2 hours for anyone else. Thanks
One potential consequence of this would be that some users (especially new ones, who are most likely to no ID and have an Unknown) might get confused or frustrated about why their observations aren’t appearing/being submitted. I think an opt-in version of this feature (like the one @egordon88 linked to above) would help avoid that potential side effect.
I believe the real solution to this is the aforementioned draft mode that has been requested repeatedly, and that would be very useful in a variety of circumstances.
later in that thread, choess suggested that “embargo” would be a good term for this kind of feature. another way to think about it would be to either publish immediately or at a required scheduled time in the future (up to some reasonable time limit).
in general, the iNat API doen’t like it if you make too many requests over a short period of time. it’s possible that if you’re going through and requesting a CV ID suggestion for a ton of observations in a very short period of time, the API is throttling your requests. if this is what’s happening, you can just wait a few seconds, and then try getting the suggestion again.
I offer a different suggestion which may not be practicable for large uploads or those with timely needs: I take the time before uploading to add as many identifications to my files as possible–at least to order, family, or genus if practicable. I’m in no rush to upload “unknowns” onto iNat–for the very reasons the OP has suggested: It makes work for others and may confuse newbies. I may upload subsets of a day’s observations periodically as I get them IDed to my own satisfaction, or I sometimes wait until I have a whole day’s batch ready for upload. I’ll actually hold back “unknowns” for some period of time until I can study them offline and offer at least a general ID for upload. The consequence for my own output is that it delays my uploads, sometimes by a regrettably long time (days, weeks, months). For me personally, I value the utility of identified uploads (for myself and others) more than the timeliness factor. But this is just my choice of work flow.
I voted for this feature request, but I’m not hopeful that it will be implemented. I’ve been wondering lately why more observers who want the extra time don’t just mark their observations “casual” temporarily, to take them out of the “Needs ID” stream. I don’t think there is much chance of anyone identifying those within the desired one or two hours. This would be a convenience to both the observers who want more time and the identifiers who would rather help those who want identifications immediately.
I like this idea though for a slightly different reason. I fairly frequently see observations uploaded from the app which seem to fall out of sync with the website. Either detaching the photo from the location/date or not synching the assigned species name correctly. Similarly, patchy internet can mean the species name are sometimes added as free text instead of assigned to a taxon.
As a result they are added as ‘unknown’. I generally catch them and amend the details but a short window (e.g. 2 hours as suggested) would be useful to avoid others unnecessarily undertaking the edits.
I can imagine a downside for this where iNaturalist is being used to support engagement events, e.g. a bioblitz, where it would be preferential to see ‘unknowns’ immediately to facilitate support and feedback to attendees.
I have noticed the “suggest ID” slows down and then freezes on “loading suggestions” after batch uploading more than 80 observations at once. I assume it’s a traffic issue. I sometimes just have to wait like 30 seconds but sometimes it’s frozen on loading for minutes on end. I will then just upload the batch out of frustration and ID the unknowns after they upload and the AI is working again. The ID suggestions helps so much for rapid uploads that you know 90% of the ID’s and can rapidly blast through them at a glance without typing anything. I can usually get the unknowns changed fast after upload, but for people who need hours to attach ID’s I can see having the request in the OP being reasonable. Sometimes the moth people have ID’s placed on my unknowns within minutes, before I can even get to them myself?
It might be due to your computer slowing down due to memory shortage, but I’m not sure. The upload page uses up a lot of RAM, I would not put more than 50 photos or so on it at the same time. I rarely put more than 20-30.
I’ll have to test that theory on another computer one day. i always thought iNat was becoming annoyed at the volume of uploads and was slowing time the more time I used the AI suggestions.
It is definitely due to load, I don’t see it as a problem, but as expected behaviour, because map won’t work on heavy load, waiting is still needed to properly prepare observations before posting, when map is working well it means cv will work too, it doesn’t stop working after 50, I upload 100+ photos and it is definitely working for all of them.
Maybe adding a notification that adding less photos at once could ne a good idea. But it’s not on website side, on any webpage if you decide to put 150 photos to upload simultaneously it will load that page a lot.
As someone who IDs unknowns, we support this or the linked-to draft mode because we often see recently uploaded batches from user with thousands of observations and we are fairly certain they are going to do more work on them.
When ID’ing unknowns, we filter for ones that have not been uploaded in the last few days: near the bottom left of the filters for identifying, we choose “more filters” and then on the bottom right we set the “Date Added” range to be days or weeks prior.
I agree with what you’ve added, @wildwestnature - I also was filtering for a week or more old to try to get past the dozens of ‘Unknowns’ from people who have made hundreds of observations.
But then I realized those newbies were being ‘ignored’ along with the batch uploaders. Just this morning, I went immediately to yesterday’s Unknowns and just mentally filtered out those chronic Unknown batch uploaders. I do filter to my state and I use the explore method to filter Unknowns so its easier to see the profile icon and sense a pattern quickly.
But the folks with brand new accounts are getting attention more quickly.
We won’t be moving forward with this. Technically it would be a lot of work (rebuilding iNat’s entire index) and it would delay IDs (and potentially needed auto obscuration) for users who aren’t purposefully adding unknowns they plan to ID later.