Also - Accuracy Not Recorded.
That could use a new DQA, or a popup before you can submit the obs.
1 Like
The main issue i have is, there is technically no correct way to make an observation casual due to accuracy range despite being able to make ridiculous ranges that wrap around the earth.
There is also a loop hole where if i set the size to the planet, i can put locations where ever i want and technically not be wrong. I could set observations at Antarctica, Madagascar, Tokyo, Argentina, The Mariana Trench, etc with accuracy circles that encompass the entire planet and not be marked location incorrect per guidelines. Even if they were all observed in Miami.
Interestingly looking on the mobile app, my dandelion observation with a 384,400,000km accuracy range appears with the accuracy circle only in the indian ocean with the pin outside of the accuracy range.
So in a strange loop hole way, it could be marked as location not accurate as the pin isnt even in the accuracy circle per mobile version. Although on desktop, no circle appears, and it also appears differently in the observation editor.
2 Likes
This reminds me of herbarium specimens I’ve seen, where there is NO locality information whatsoever. Perhaps the collector simply wanted a specimen for their own use, but as a specimen in a public herbarium, it is virtually useless to researchers. One could argue that the herbarium curators should discard all such specimens (and I’ve seen that happen sometimes), but sometimes such specimens are useful as teaching specimens (which get handled a lot and thus may get damaged) or useful as part of a collection from a prolific botanist. Most curators have so much to do that weeding out specimens with no locality or a very broad locality (“New England”) is a task they never get to.
5 Likes
The main question I have is, what is the benefit of making the observation casual?
If the spatial data is set out correctly, even though it might encompass a very large area, or even the whole planet, a researcher planning to use that data has all they need to make an informed decision whether this record is useful to their research or not.
Making these observations casual has some clear disadvantages, primarily that this record will be hidden from potential identifiers, but what advantage exactly does it provide?
8 Likes
Your eliminating an observation with no location data. An observation with no location data can be assumed to be on earth. An observation where the accuracy circle covers the whole planet is the same, its somewhere on earth with no further clarification. You can also bring up observations that are private, they are somewhere on earth, but can get RG but arent exported i think for research?
1 Like
Wait till mankind conquers new planets replete with extraterrestrial lifeforms, and intergalactic citizen science becomes a thing.
Then, these old obs accurately located ‘from Planet Earth’ will take a new meaning… ;)
(I agree with you, in our current state of affairs a location accurately described as ‘somewhere on Planet Earth’ is as useful as no location. That’s why I think it might be reasonable to place the upper boundary of ‘location’ at, let’s say, the hemisphere.)
3 Likes
- It gets it off the map. A dot that could theoretically be from anywhere on earth should not be mapped to a specific place on earth. Having accurate range maps is good.
- It helps keep identifiers from wasting time. If I’m identifying observations from California, I don’t need observations from Florida randomly popping in and confusing things.
- It encourages useful contributions to the project. We want people to have observations with good locality data. This should be actively encouraged. It’s the same reason we have any other standard for Research Grade (e.g. no captive animals, must provide evidence of organism, ect
3 Likes
I can only partially agree:
Yes, I suppose we deal with this in my workplace by not putting any dots on the map for records with a spatial uncertainty of more than 25 km. Just doing this would probably be straightforward and not have the drawbacks of casual grade. Another option is to give records above a chosen uncertainty radius a different icon (as we do with obscured data).
You are jumping to the assumption that identifiers are uninterested in those records, and also would be taking away their ability to make that decision themselves by removing the record from consideration by default. This results in a moderately useful record becoming a fairly useless record, as it will most likely never get a supporting ID. Spatial data is only a subset of the potentially useful data that a record provides.
This assumes that the lack of spatial precision is a choice of the observer (not always), and also assumes that a project does not have the capability to omit records of less than a certain precision (they do).
1 Like
If I am filtering for observations from California, I think it is safe to assume that I am only interested in identifying observations from California.
I am not assuming that every specimen that lacks locality data is due to a choice of the observer, only that a non-zero amount of them are. It’s the same argument you could make about other criteria for Research Grade, such as date and evidence of organism. Should we also not require date or evidence as well?
1 Like
I would think most people looking to exclude low precision observations would also want to exclude those, but I can only really speak for myself.
1 Like
If people actually want to identify Casual status observations, they know where to find them. I’ve never found the “but they rarely get identified” argument persuasive against giving a record Casual status.
If casual observations aren’t getting identified, that’s probably because…identifiers don’t want to spend time identifying them. And we shouldn’t use the website UI to push people into looking at observations they otherwise wouldn’t want to be bothered with identifying.
1 Like
I don’t love the idea of making records with very large accuracy circles casual (it will just be another ‘hidden’ casualization that makes people resentful), but maybe it would be nice to have a hard-coded upper limit to the size of accuracy circles one can draw, even if it’s quite large.
There are currently ~220,000 verifiable observations worldwide with an accuracy circle larger than 1,000 km (and ~4.4 million with a circle larger than 10 km, which is more than I expected). When an uploader manually types in my city into the ‘location’ box when uploading an observation, it makes an accuracy circle of ~15 km (and always puts the pin in one particular traffic circle, which looks like a real biodiversity hotspot at first glance!).
I do sometimes find it frustrating to find, for example, a map pin for an uncommon woodland species sitting in the middle of a very urban area, which is technically potentially correct since it has a very large accuracy circle. I like the idea, suggested above, to
2 Likes
It would be nice if more inatters added location notes in their records, such as a state park name or even a landscape feature, to supplement the coordinates provided. Coordinates are great when accurate and precise but when they’re wrong or imprecise the location notes can be very useful to clarify a location or indicate when there’s an error.
4 Likes
Then another popup - please check if you agree with the location as named?
No?
You can edit it to suit then. Thank you.
Just for some context, this scenario is a problem of vanishingly small proportions:
There are a grand total of 920 observations (out of >248 million) on iNat with an accuracy value >24,000km.
2 Likes
There is none.
It is up to the researcher to decide what accuracy is useful for their project, not ours.
4 Likes
“Research Grade” doesn’t mean the observation is perfect. It means it has a photo or recording, a location, a date, and 2+ agreeing identifications. That’s all. Researchers using the data have to inspect it and see if it meets their needs. Researchers expect this, learn to deal with this for all data. (Maybe they don’t expect it the first time they try an analysis, but they learn.)
It would be inappropriate for me to to decide what limits everyone should put on “useful data.” Maybe for my work, an accuracy circle bigger than 0.5 km makes the data useless. Does that mean it won’t work for you? How would you feel if you learned that you were working with only half the data points for your rare plant because somebody had stripped out all the observations lacking accuracy circles and half the observation had been made by a person who didn’t understand or use accuracy circles? I would be furious, myself.
Keep in mind, too, that in many cases we can interpret well enough observations that have large or no accuracy circles. (We who work with herbarium specimens have to do this all the time. I’m sure you can learn to do it, too.) If the location of a terrestrial plant is in the middle of the lake, it was seen along the shore. If it’s at park headquarters but it can’t possibly be there, it’s somewhere in the park. And if there is a latilong but the accuracy circle is bigger than the earth, the organism was somewhere near that latilong point and the observer doesn’t know how to use the accuracy circle. For example, if the latilong is in Florida but the accuracy circle is big enough that the observations might show up in California, the observation was in Florida.
Maybe that data point for a tree in the middle of the lake isn’t good enough for you because you need to know where along the shore that tree was? OK, don’t use it. But don’t decide whether I’ll find it useful! Especially, don’t make the decision for me just because you are annoyed about having to deal with the sloppiness that is an inevitable component of citizen science.
10 Likes
The reason we don’t identify wild plants put in “Casual” is that to find them we have to wade through all the observations that lack dates, lack location (completely lack location!), lack photos, are copyright violations, are observations of humans, and are captive/cultivated (which should be a separate dimension, but that’s a separate topic). Any good stuff there is buried under piles of trash. We should be very, very leery of putting any observations of wild organisms into the trash heap that is “Casual.”
4 Likes
that would be a typical farm dam here. Where the tree was drowned when the dam was built.
1 Like
True! (I was thinking of an older lake, like Lake Tahoe, but you make a good point.)
2 Likes