Hey, could someone contact the project manager ASAP? The project settings don’t include research grade, so all the wild confirmed observations are getting kicked out.
Do you know if there’s a way to filter this page? If not, do you think that would help?
I don’t think so. I also think the atlases are pretty sparsely populated in Asia
I haven’t figured out a way to filter, either. In situations like this, though, what’s most critical is having atlases for the common, often-identified species that will get suggested by the AI, so that eventually the identifier could be warned that they’re selecting something known to be out-of-range.
Is there a movement towards that? Long ago when i was helping set up atlases i was told they were mostly for taxa split and adding range of things wasn’t that helpful. I’d love it if that wasn’t the case any more but maybe we should get some admin guidance or something.
I feel like a lot of the points raised here have already been raised multiple times in the Google Group in the past. I’m not sure how helpful it is to bring the old threads back up, but I’m also not sure how productive it is to rant again and again about the same issues.
Here are some threads I found quickly: 1, 2, 3, 4, 5
There are a bunch of separate issues that all contribute to the frustration, for most of which there is a related suggestion on the forum: students motivated by grades, duplicates, multiple photos of the same organism uploaded as separate observations (combining photos during submission is not super intuitive in the app or the website, but the new tool mentioned here will hopefully help a bit for after the fact), the user not adding any ID, and large amounts of (mostly) useless pet and potted plant observations. I think the last one is probably the hardest to deal with.
I think perhaps the best move on this new forum is to find a feature request that matches what you think we need, or create a new one, write up your concerns and vote for them. There’s no other way to get things over from the old google group and it seems like this forum is easier for admins to see and interact with us.If it gets a lot of votes and comments, we are more likely to get a discussion that includes the admins as well. I know they are aware of these issues, but as for their strategies and ideas, only they can speak for those.
There’s already a feature request about student accounts, though other ideas might work too. There may be one about flagging duplicates too. The last concern you mention probably relates more to what sort of outreach and marketing for iNat occurs, or else how people find it organically on their own. i personally think the site should do more to attract ‘power users’ and ID helpers and not try to recruit any more students, etc. But i don’t know if that counts as a feature request or not. And i’m sure others may disagree.
Oh, I have no idea what staff is planning. But looking at Tony’s next-to-last bullet point after the retreat, “Continue to improve the Computer Vision model and the use of geospatial data in suggestions,” I assume something is planned. It could be just based on geographic proximity, but atlases seem like the natural mechanism for capturing that kind of advisory geospatial data. It would have to be crowdsourced, since there’s no way staff alone could keep up with maintaining taxon distributions.
I have no inside knowledge, and it links into this : https://forum.inaturalist.org/t/what-should-represent-a-range-for-a-species-in-inaturalist/399/6
But thinking about it from a technical perspective (or if I was tasked with coding this), atlases are the last choice of all the available options that I would use to determine geographic proximity for the computer vision.
I agree with both comments. I think the system is working as designed, but is producing results that are considered unsatisfactory. I am not a huge believer in changing human nature in the short-term, so I think changes to how the system works would be more effective, but I’m ready to vote on discrete issues in Feature Requests.
i understand all of your frustrations though, i feel that way too
I’ll take your word for that, but there’s also the flip side of the puzzle to consider–how do you successfully crowdsource and maintain distributions for every taxon on earth that people can observe in the field? Not that the current atlas system is the last word in representing range data, but that’s an important social consideration, and the atlases (for good or for ill) imitate the typical way we currently get range data (at least for plants).
this comment on the Penang project page:
@lingeshwarry what a nightmare, as useful and needed for iNaturalist community as two broken legs! I wished you got the meaning why, still won’t try to explain further.
Posted by erwin_pteridophilos about 11 hours ago
I have messaged him directly and asked him to consider deleting it.
I agree. When we all joined we were one person learning the ropes, and there were dozens of helpers to guide us. This is 1400 joining, and still only the dozens of helpers, so they don’t have the same level of 1:1 learning that we got. Also, they are joining and hitting the ground running, 200-700 observations in their first day or two! I don’t know about everyone else, but I learnt much slower, a few observations at a time and had the opportunity to alter behaviour before there was a huge impact.
I reckon we need an intro video, translated to many languages, that cover all the basic learning required to be a “good community participant”, and have all new users required to watch that before their account is activated. Re-watching it could also be a requirement for reactivation of a suspended account…
The problem is 95% or more of these folks will never open the app again once the Nature Challenge is over. Just look at how many observations came in from Malaysia today after it was over.
All the more reason to get them schooled up on being a good observer BEFORE they start!
Unfortunately their motivation is not to be good observers, it is to record as much as possible as quickly as possible given the timeframe of the Nature Challenge.
I’m kind of dreading the City Nature Challenge after this debacle. I don’t mind newbies throwing in a few cultivars when starting out, but it’s usually just a few!
I’m almost thinking we need some kind of barrier to entry before allowing users full site use - something like successfully IDing X number of their own observations before they can post more. I don’t know if that’d work though.
I felt that way after last year’s City Nature Challenge. But this year for the first time my area has one, so I will participate.
I agree. I have been trying unsuccessfully to make a Feature Request on limiting the ability to turn something Research Grade. Your idea would expand that, and I like it!
It has occurred to me that perhaps a simple solution to this kind of mess would be to simply disable the auto-suggest ID for regions that have limited numbers of observations (until there is a critical mass of observations for it to work with, whatever number that may be). It seems like it’s rarely helpful for finding a correct ID in these areas currently, and just encourages people to pick a suggestion at random.