As has been said already, huge accuracy circles are a vanishingly small problem.
1 Like
I am mostly concerned with observations of potentially/likely cultivated ornamental plants that are often made in urbanized areas. It happens that just one photo is uploaded with no depiction of the environment where the plant grows and the position accuracy does not allow to be sure if the plants has been observed in a private garden/managed green space or has escaped from cultivation. Unfortunately this may happen rather frequently.
3 Likes
My opinion is when you run across this problem, message the observer.
Anyone using iNaturalist data for research has their own obligation to assess the information.
From what I can gather from this thread it is a minor issue at best.
1 Like
I post lots of plants in urban areas and I can tell you that the accuracy circle almost never allows you to determine if the observation was cultivated or wild! Even plants on private land, even plants growing near a house, may be wild. The cultivated / wild problem cannot be solved through improved accuracy circles.
6 Likes
Sometimes I did so and in few cases the issue has been solved. But, have you ever tried to do so whenever you find that issue? Think about duress users or other users who are here because they earn money with iNat (e.g. some educators). Do you think they are interested in making good quality OBs?
Yes, the problem cannot be solved so but in some cases it could be improved.
1 Like
I still regret that iNaturalist,org does not allow to use other geometrical figures than circles. And I still don´t undestand: Why?. A river and its riverside can be a habitat two kilometers long and 20 to 50 meters wide. Behind this there can be meadows with completely other plants, fungi and animals. But you can only create mixet entities. Not very satisfying!
2 Likes
Because with coordinates plus a single number (radius of the precision circle) you can define the location. To define more complex shapes, you would need to save much more information (for example the vertices of a polygon) and it is unclear how this information would be used. You can always use additional notes to explain locations, for example, it was within 50 m of River X between points A and B.
1 Like
I think the overwhelming majority of observers do in fact make good quality observations.
I believe that whenever you or any other person using iNaturalist feels under duress that person needs to take a break.
I think the wisdom is whoever is trying to use the data can skip what does not fit their criteria, not put pressure on themselves using unsuitable information.
You do know the first co-reason of developing iNaturalist is to get regular people to engage with nature. We are not here to further your projects or careers.
Now, if this is something you want to pursue, you can make it a feature request, and offering appropriate parameters of solution and see if it flies.
4 Likes
Nope. I can’t agree.
It’s since my first visit in the forum that I read this concept repeated over and over again. Nothing against tne engagement of people with nature, indeed… but “people”, especially students, are usually enough smart to understand the use of a box in an app and the reasons why it should be used. But explaining things can be time consuming and laborious…
I am not referring to you, but I am tempted to think that always justifying the misuse of iNat is somehow disrespectful.
iNaturalist is an online social network of people sharing biodiversity information to help each other learn about nature
2 Likes
Duress users usually means students who have been assigned to use iNaturalist. They cant “take a break” if they want a passing grade, and if they aren’t into iNaturalist of their own accord, why would they put in more than the minimum of effort to pass the assignment? Responding to a message about their accuracy circle probably isn’t in the rubric.
5 Likes
Make it a feature request.
“Think about duress users or other users who are here because they earn money with iNat (e.g. some educators)”
Nothing about duress of students.
No one needs to agree with my opinions or thoughts. They are just as legitimate as anyone else’s.
No one is saying that someone shouldn’t participate. No one is saying that their observations should be removed. No one is saying Causal observations are worthless.
What we are saying is there are existing, already existing, criteria for when an observation can count as “Research Grade”. Some observations are eligible for Research Grade, and others aren’t. You don’t need “Research Grade” observations to participate in iNaturalist. And a locality requirement is already part of the criteria for Research Grade. It’s not something new, radical, or exclusionary.
2 Likes
Please tell this to all the new users creating sock puppets and faking DQA to push their obs into RG.
1 Like
Possibly something has been lost in translantion but I thought I had written in plain words. maybe there would be no nocessity to lecture with the link.
Instead, I wonder what is a true engagement with nature. Is it mandating students to make observations with blurry photos that are often wrongly identified by the CV? Is it unleashing people who know nothing about iNat telling them that the the more observations they make, the better it is?
4 Likes
I would like 3 popups from iNat.
Say after 100 obs, then 500, then 1K.
‘Welcome. Thank you now … here’s … how you can improve your obs by … how to ID … join and support a project since you are interested in … how to sort out your problems - mavericks, disagreements, Geomodel Anomalies …’
I am fine with Newbies floundering, we all had to learn iNat guidelines. But when they have thousands of obs - should do better!
3 Likes
Of course, we should be patient. In 99,999…% of cases they are just misinformed.
I remember that many suggestions have been made through the years to address a number of issues. Some still seem not to be implemented.
1 Like
I would say that Casual observations are very nearly useless. Any potentially useful ones in Casual are buried in the many, many truly useless observations and cultivated plants. We should not put potentially useful observations there, like ones that are otherwise OK but have ended up with very large accuracy circles by accident.
3 Likes
There’s no DQA item that relates to this (nor should there be). The “Location is accurate” item has nothing to do with accuracy circles per se.
The only RG requirement for the location is a valid lat/long. The accuracy circle and place-name are entirely optional qualifiers with no fixed interpretation. They may be used to judge whether the lat/long represents the true location with sufficient precision, but their usefulness with respect to this does not in itself affect the RG status. If the true location can be inferred from other observation data, the accuracy circle can act as a visual guide in assessing whether the lat/long is close enough to its intended target (i.e. whether the location is “accurate”). As we all know, it’s often too big for this particular purpose, but that is not in itself a reason to downgrade the RG status or make the observation casual. There’s no more reason to do this than reject an observation because some of its photos are out of focus.
The accuracy circle can also be interpreted by data consumers according to their own arbitrary criteria. Typically, it’s assumed the value is a reliable, independent guide to the true precision of the lat/long - but in practice, this is rarely ever true, even when the data comes directly from a recording device. The notion that iNaturalist should enshrine this sort of interpretation in the DQA is at the very least highly questionable.
There’s no doubt that the general quality of location data is in need of improvement. I suspect that the accuracy circle is often singled out in this regard because it’s just a simple numerical value and it seems like there should be an equally simple way way to leverage it to make improvements. This is essentially what many data consumers try to do - it’s so tempting to reduce it to a binary value and treat it like a tick-box that can be processed by a crude batch process. By comparison, verifying the lat/long by inferring the true location will almost always require time-consuming case-by-case assessment by humans. And of course, the other side of the coin looks much the same. Creating (or faking) an acceptable accuracy circle is trivial, but far fewer users will go to the trouble of working out a precise place-name that convincingly matches the lat/long. Persuading more users to do the latter isn’t easy, but it’s undoubtably the best way for iNaturalist to try to make improvements.
7 Likes
There is an assumption here that the center point of most observations with very large accuracy circles is relatively close to where the observation took place, and is a result of technical glitches with the GPS of people’s phones/cameras. While I think that’s certainly true for some of them, there are also many records that get large accuracy circles because people manually type in very large locations (e.g. countries or states) when they upload photos without location data.
If you type in ‘California’ as the observation location when uploading on the web browser, for example, it will make an accuracy circle ~740km wide, and place the pin in the central valley, and if you search that location, you’ll find 780 records (accuracy over 700km) with a wealth of sea lions and sea birds and other such things. Same goes if you type in a country, like South Africa (333 records) or India (600 records).
For these records, the pin/center point is unlikely to have any correlation with where the observation took place (though of course the observation presumably took place somewhere within the accuracy circle). Not saying those records are necessarily problematic or anything—just wanted to point out that the lat/long for such records probably isn’t the best indicator of the true observation location.
3 Likes