How much time passed since observation affects on a chance to get ID?

I have checked FAQ and didn’t find an answer. Me and many other people notice that often when there is no identification for the observation, it helps to delete an observation and upload it again.

Could anyone explain to us, is it really works, and time passed since really matter? And how it is implemented in code? Or it just happenstance?

1 Like

By default, the Identify page loads the most recently-uploaded observations that are “Needs ID”. So yes, it’s biased towards observations that have been recently uploaded. But some identifiers do go through older observations - I’ve been doing that lately.


posting exactly the same thing again seems dysfunctional to me. maybe you can get someone new looking at it, but just as likely, the same people who looked at it before and chose not to identify for whatever reason will see it again and still can’t or won’t identify. (that would be wasting those people’s time looking at the same thing again, since they wouldn’t have the benefit of having filtered out such an observation as previously reviewed.)

i think the best way to get an identification is to identify the organism as best as you can, and then someone else will be more likely to come along to refine, confirm, or suggest a different identification.

there is a graph that shows average time to ID here: note that these are averages, and there are many factors that affect whether an observation can be identified.


Me too - I often start with the oldest first for my area when I’m identifying things. However, a lot of other people will start at the newest uploads, so over time I think the chances of having someone look at an observation and add an identification might decrease. Type of organism also affects this (for example birds often get IDs very quickly, while plants don’t).

1 Like

I see your point and agree, mostly I don’t do it.

But the question was about an algorithm of feed creation, technical question.

For the sake of integrity I would suggest you do not reupload things. Some people do that to avoid IDs they disagree with and in my identifying experience it is a bit frustrating to see repeats–especially having to go and check if something was stolen, deleted, or what else happened.

Unfortunately just due to the mass number of observations, many will remain unidentified for long periods of time. You can always tag identifiers or send them a DM if you want to personally request they take a look. Certain groups are also just super tough to ID or have very few active members who focus on them.

As was said already, the Identify tab by default just filters the latest uploads (by upload date, not observed), but once people review observations they get taken off the list for whoever left an ID, so they end up seeing older stuff, varying by how often the taxa they are looking at get uploaded; there is no sort of algorithm biasing certain people’s observations–it’s all based on filters. I would mess around with the tab yourself so you can get a feel for it.


As I mentioned, it was more technical question because some of us arguing is the time matters or not.

I agree with you about reuploading

1 Like

it’s not really a technical / algorithm question though. everyone interacts with the system differently, and the real answer to your question depends on how the potential identifiers of your particular observation interact with the system on any given day. and even if time matters, as mentioned previously, there are other factors that likely matter much more. if you want specific statistics about which factors are most important, you should do some research on that topic because i doubt anyone has ever done a systematic study on that topic before.

that said, the best way to get an appreciation for how the identification process works – better than can be explained in the forum – is to do some identifying yourself.

looking at your observations at a high level, it doesn’t look to me like your identifications are getting identified any less than typical.



50% of IDs come from just 545 identifiers! We are the top 0.3% of iNatters (I wonder what percentage that is of ‘still here after the first year’?) Imagine the fallout if one of those accounts is deleted.

If. You want more IDs on your obs? Please reach out and help to ID whatever you can!

1 Like

I personally do not need IDs often, if I can get the name of species from the neural network. Because I always google animal/plant/fungi by myself after I’ve got the name.
So I don’t complain I just wondering) And wanted our conversation with my peers be more meaningful then just sharing our experience with each other.

Before I asked here I visited GitHub with the code but there was no documentation I could find and I am not familiar with Ruby on Rail. So that was main reason I asked here.

Thank you for your answer and the links, I am looking at them right now!
OMG these Sankey diagrams are gorgeous! Thanx for that either!

Thank you for your answer, but it was not my request)

Of course the labour of identifiers is so great and I appreciate it much!

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.