Pictures of newly uploaded observations showing up with delay

Please fill out the following sections to the best of your ability, it will help us investigate bugs if we have this information at the outset. Screenshots are especially helpful, so please provide those if you can.

Platform (Website)

Browser, if a website issue (Firefox)

URLs (aka web addresses) of any relevant observations or pages: e.g. https://www.inaturalist.org/observations/219459905

Screenshots of what you are seeing (instructions for taking a screenshot on computers and mobile devices: https://www.take-a-screenshot.org/):

This is what it looks like on my dashboard:

Description of problem (please provide a set of steps we can use to replicate the issue, and make as many as you need.):

Something that I have been noticing regularly lately is that new observations are not displaying images (see example above). Sometimes, users forget to upload images but I don’t think this is the case with most of these observations. The images show up later with a delay of several days but the picture does not display initially, neither in notifications on my dashboard, nor on the observation page itself. It is now happening far too often. Anyone else having this issue?

Thanks, Matthias

This is what the observation page looks like:

The contributor used computer vision to identify the species but the image is missing. An image was obviously uploaded by the user or else CV would not have been able to suggest an ID. Based on the time stamp, I visited the observation page 21 minutes after the CV identification was made.

Now over two hours that the observation was created and the ID was made, but the image hasn’t loaded yet.

Tho photo could have been removed by the observer before the upload.

Leave a comment on the obs, for the observer.
Your variation of
No photo
In the observer’s interest to sort it out - as no photo pushes it to Casual.

This is not a problem that is connected to any observer. It happens almost every day, sometimes more than once. It appears to be a software issue where observers upload images but for some reason they appear on the website with a delay of one or more days.

Computer vision would not have suggested “Astragalus pectinatus” if the observation had been created without an image.

This happens regularly. I see several every day.
Invariably with one of the apps.

Simply, if the app is synced, but for some reason the signal is compromised, then the observation uploads without the picture. The problem is simply solved by syncing.
However, some observers dont notice that their uploads did not complete, and only revisit some time later.

The AI CV makes IDs because it uses a smaller picture: not the original. Often the IDs are made long before pictures are uploaded, and IDs are also possible in the field, even if one only uploads the observations until much later. So a CV ID is unrelated to this issue.

I just leave a note.
“Please sync to reload your pictures”

2 Likes

I am wondering: Could the app be modified so that it requires the image to upload in order to create the observation? Why have an ID and not provide the image it is based on? Do we really want observations without images?

Here is the next example. I get one or more every day:

1 Like

I’ve recently started adding comments to photos without observations, along the lines of “Hey, you didn’t upload any photos. Was that intentional?”.

I’ve gotten one (paraphrased) “Yeah, the app upload is weird when I’m in the field with a bad/slow connection. The photos will show up on their own later.”, and a handful of (again, paraphrased) “I see it on my end, I don’t know what your problem is, bruh”, and then by the time I get that comment the photos are now present. But I’ll swear to you that the observation had no photo and was casual when I left my comment. (I’ve gotten other replies too, like “Oh, hadn’t noticed, thanks!”, but those aren’t relevant to the context of this topic.) It’s also not uncommon that no photo will ever show up at all though.

I haven’t noticed where there’s a CV ID but no photo until now, but I went back and checked one of my ID queries and a handful of the “ID’ed but no photos” observations have the computer vision icon next to the ID. Previously I just marked these as reviewed and moved on, but now that I know this is a thing, I’ll probably go back and add a comment asking the observer if they realized that there are no photos attached to the observation. (6/20/24 EDIT: I did go back. There were 8 of them total. 1 was from the iPhone app, the rest were from Seek. Collectively, the 8 had submitted/uploaded dates ranging from 2020 to 2023.)

I’ve actually been thinking about this the past few days and thinking about creating a post about it and then I stumbled across this thread here. It’s really hard to tell when someone intentionally omitted a picture versus when it’s due to some other factor (slow internet, a mistake, etc.).

Not directly, but I see evidence of it. Doing IDs for Puerto Rico, I frequently come across a copypasta by a well-known, active identifier commenting “No image?” The observation has an image. Until you brought this up, I wondered why so many observers in Puerto Rico would forget to upload images.

Perhaps related to this, but observations made days or weeks ago but without photos are appearing completely blank right now

See:
https://www.inaturalist.org/observations/222660527
https://www.inaturalist.org/observations/217551526
https://www.inaturalist.org/observations/217038272

1 Like

Not directly related, in cause at least. This (photos showing up later than observation) has been a thing for a time period measured in whole years. The “can’t open observations without photos” issue started a few days ago (see https://forum.inaturalist.org/t/can-t-load-observations-that-lack-photos/52645).

It is indirectly related though in that now these observations with a photo not yet shown up can’t be opened either. (Though they can still be surfaced through, and IDed/commented on, through the Identify page.)

EDIT: For some reason I was conflating this bug report you and I are conversing under with https://forum.inaturalist.org/t/suggest-id-infinite-loop/52506. :man_facepalming:t3: Though the takeaway still holds true. I’ve personally noticed the delayed photos for several weeks. It may have been happening for longer (some of the older observations with no photo but a CV suggestion seem to indicate that is a possibility), though I myself only started paying attention to it a few weeks ago.

If you can send some URLs of these to help@inat it would be great. Then we might be able to look in our logs and see exactly what happened and tease out any patterns.

Do you want examples of observations without photo and having a CV-aided ID suggested? Or are observations without photo and no ID at all (but which mysteriously get a photo later, with or without comment from the observer) also helpful? (Or both?)

Is there a certain timeframe that would be unhelpful? For example, is an example of an observation uploaded 3 years ago helpful? Or do you need relatively recent examples?

I think that’d be best, then we know there was a photo at some point.

Within the last month or so, please. Thanks!

@tiwane I’ve just sent you (via help at inaturalist dot org) 12 examples from this month just from my state.

For everyone else, what I did was queried for observations created on/after June 26, 2024, with an identification, without photos, without sounds, in Nebraska USA, and of any quality grade: https://www.inaturalist.org/observations?created_d1=2024-05-26&identified=true&photos=false&place_id=3&sounds=false&verifiable=any. My query returned 82 observations. (Without date criteria it returns over 5600 observations.)

I then selected the 1st observation in the list, as well as the 2nd observation in the list, and then every 4th observation after that until I had opened 11 observations. All but one of them had a CV-aided ID but no photo. Most were uploaded via iPhone or Seek, so I modified my query to only return Android-uploaded observations and found that it too returned observations.

I mention this in case anyone has an idea to check using some other methodology. Based on what I checked though, it seems that this is a pretty prevalent thing.

All of the observations I checked had a matching date and time for when the observation and its ID were created. Presumably the app is uploading the observation first and then uploading the photo when it can? I’ve had other users anecdotally tell me this is what happens to them frequently. I can try to track down who told me that if that’s helpful.

1 Like

Thanks for the examples! I’ll bring them to our team.

1 Like

So what we think is happening is that the observations are being uploaded in an area with unstable cell signals. The non-photo part of the observation can be uploaded easily but the photo fails, leaving observations like this. What we can’t determin so far is whether the user sees that the observation still needs to be uploaded, or if it looks OK on their end. I’ve asked a few people but haven’t gotten specific answers.

1 Like

Here’s a newer example where the user elaborates on what it looked like on their end: https://www.inaturalist.org/observations/237695988

Interesting, thanks.