And that is precisely what many people do, don’t add ids. Can’t say it’s the best way of things, but possible.
With that data, it’s cool enough it exists at all, in your case I would exclude in url a taxon of domestic corellas, then id all that left, with 429 verifiable records of this species, it will take an hour at max to check all and id, to make process faster, I’d ask a fellow scientist to go after me and also apply domestic id where mine was the first, that way search will show only wild ones pretty fast.
There’s no opposition, and if iNat will add it, it will be used, it’s just even now it’s not too hard to separate the two and scientists have to do a lot of work when they use such data anyway, there should be no expectation it’s all clear from the start. Option now doesn’t exists because iNat needs a lot of work and group working on it is pretty small.
It was clearly a priority, because 244 posts and 18 months ago @ loarie proposed a very clear method for something like this that still hasn’t been implemented because of opposition here.
In the grand scheme of things, it’s quite a small deal, but it just confuses my why so long after a (at least partial) solution was proposed there is still so many people who don’t want to see something that would improve the quality of iNaturalists’ data exports.
These observations also have the significant potential to influence iNaturalist’s machine-learning methods, with escapee observations potentially making it so that the machine-learning suggestions begin to suggest species that are far out of range, because of these records making it to research-grade and being unable to be flagged through data quality attributions.
If it was a small deal, it’d be here already, we’re waiting fow new explore and new notification system for how long? Or check annotations’ topic, that was started in 2019 and how much of proposed annotations were added by this point? iNat is not a global corporation and it grew too fast to make such big changes in a matter of months, there’s an ongoing battle to keep speed of website up, likely a real priority of developers.
One possible and I think workable solution that you allude to here Marina could be to establish domestic variety taxa, similar to what iNaturalist has with ducks, geese and even Cockatiels, and encourage users to identify clearly escapee individuals to this variant. This would allow the easy filtering of these observations by any researchers. The only issue is that would require a fair bit of upkeep by the community taxa curators, but I think that can be managed without intervention by the already swamped (as you’ve pointed out) iNaturalist staff.
I am confident opposition here is not at all the reason why this has not been implemented. It’s just a matter of how imporant iNat feels the issue is vs how little resources and staff time they have to go around. After all we have been cheering on the notification system revamp for 3 years now and that hasn’t happened any faster.
We don’t know what scientists will want in the future. It’s quite possible they’ll never need more records of Douglas-Fir from Oregon, where it’s very common and there are other ways to find out where it is. But we keep posting them.
We do know what scientists need: verifiable records with accurate location and date. Let’s concentrate on providing that. Let the hypothetical future scientists worry about getting that they want from the data iNaturalist has provided.
Those data may have current or future applications for understanding the establishment of introduced populations. Given that iNat data are presence-only and don’t include most measures of effort (which makes any assessment of density, population size, or trends nearly impossible), those data on occurrence are probably one of the more useful aspects of the dataset.
The filtering you are describing really isn’t much of a burden (any of us who use community science data are always doing some sort of filtering and quality control). Always better to have maximal data available and let the researcher decide how to subset it for their particular question(s).
eBird takes almost the opposite approach to iNat and aggressively curates and removes many introduced species from their exports. In practice, the decision of whether or not individual records are invalidated as “non-established” is inconsistently applied, even when you have those decisions implemented by just a few dozen reviewers (rather than by individual users in iNat).
The result is actually more work for scientists using the dataset. For example, I published some analyses on the current and potential establishment of both Egyptian Goose and Agapornis in California. For both of these taxa, the records I was looking for were split across two data exports (validated and invalidated) which I had to reconcile. It would have saved me a step if all the confirmed observations were treated the same and I could implement whatever filtering I wanted to impose from there.
Absolutely.
@nrg800 I would suggest reading through some of the comments above from other practicing scientists who advocate similarly. Two key points:
-
Cleaning/filtering data for scientific analyses is a given - it’s not more work to deal with these data in most cases. On the flip side, inconsistent filtering of data before it gets to a scientist is waaaay harder to correct/compensate for (if this is even possible; sometimes it may not be).
-
This is not correct:
There are many useful applications for these data (I’ve personally used them). And as many others have noted, it’s really difficult to predict what will be useful to future scientists. Assuming we can predict usefulness is a losing game.
I just want to note that site staff clarified here that escaped pets should be considered wild.