Isn’t this just beautyful? Hanging around with your family for 11 years before being send off to RG- heaven? Made my IdentiFriday today 
12 Likes
I don’t know if you had the help needed yet, but I’m fluent in French if that’s not the case !
Thanks, I’ll tag you on the observation in question
1 Like
I’m not sure if it’s only the genus I identify the most or a global trend, but has anyone noticed an higher amount of misidentifications compared to last year ? CV seems to work fine on my end so it doesn’t seem to be the culprit. I’m not sure if it’s because there is more new users less used to recognize plants, if the increasing amount of observations leads to local identifiers not finding the obs before I do or something else.
2 Likes
Well - the ‘winner’ is losing sp (another 20 since yesterday) each day when I check. With 1 million obs in one CNC day - the pressure is on to ‘add more sp - not found here? - whatever’ But then I like to check our local Geomodel Anomalies, and to catch unexpected outliers when I use a distribution map.
On the leaderboard, by sp count, San Antonio edges closer to the ‘true’ winner. 4867 vs 5091. Give the taxon specialists another 2 weeks …
1 Like
I’m not noticing any higher number of misidentifications. I identify mostly in New England in the northeastern US and mostly (but not entirely) work on plants.
i haven’t noticed anything unusual, but it might be worth noting that i think that the way the new app works, if it’s not connected to the server, it’ll fall back to a simplified local version of the computer vision model, similar to the way Seek works. i’ve seen some folks comment in the past about the quality of the identifications coming from Seek. so it’s possible that something like that is happening with the new app sometimes, too.
Under 26,000 now! If you like sorting invertebrates, please join in. You can run the CV on a lot of Odonata for easy answers, since it’s learned many more species in 5+ years. You could also use the new disagreements filter to hunt for “stuck” observations.
2 Likes
FYI, we have a new filter
11 Likes
Oh I love this! Thanks for making me aware!
1 Like
This is going to be a game changer!
6 Likes
A week later the difference is about 100 sp …
4965 (still trickling down each day) versus 4873 (holding almost steady) for San Antonio
Today - the winner is San Antonio with 4830.
Cochabamba sliding back at 4686.
Graz moving towards second (thanks to the ‘sliding back’) at 4537.
5 Likes
It may say something conceptually about CNC that the rankings after the fact change only by subtracting species, not by adding newly identified ones. This could be construed as saying that everywhere is less biodiverse than people think it is.
1 Like
I think more - that well curated and managed projects reach a stable number of species fairly soon - but still they shift a little day by day as taxon specialists continue to sort thru them.
Apart from Graz at 58 % - which is unlikely to move much in future?
Most of the projects are around 40% - so almost two thirds - could still contain more and interesting sp (but also the Not Wild, or frankly broken to weed out)
The new Disagreements filter should still retrieve some good sp.
This reserve was a blank spot on iNat before (oops) GSB.
Now has 137 sp.
2 Likes
If you ID for your preferred location - and you use the Geomodel Anomaly to check for problems - the latest update is a HUGE jump!
2 Likes
It’s Friday eve and I passed 5,000 IDs in family Ripiphoridae and my saved search has only 725 left to review. Then I may need to go back through and bump some at genus and complex to RG and wait for GBIF to update so I can download the dataset.
436 records include host plant information, which is essential for my research: https://www.inaturalist.org/observations?taxon_id=453733&verifiable=any&field:Name%20of%20Associated%20Plant=
That started with a forum post 2 years ago: https://forum.inaturalist.org/t/recruiting-plant-identifiers-for-beetle-project/44430
13 Likes
With the help of a fellow IDer I found myself a new hobby while chipping away on those north american lycosoidea - training the CV with new species. Meaning I am going on a hunt to find those observations to get to 100, specifically tweaking my filters in certain ways to find them.
Other then that I skip around different settings and taxa and sometimes also continents (but most of the time I am in the US and Canada now) to keep it interesting.. however, my idea from earlier this year to finally get a bit into australian spider fauna kind of died apparently 
6 Likes
Recently I’ve been IDing rabbits and hares with disagreements at family level in North America (i.e. confusion between cottontails, hares, and European rabbits). 35 pages reviewed and still needing IDs. 21 pages left to review.
8 Likes
I’ve set out to clear the Sedum genus from the identify section. There were only 36 pages but it’s taken me ages to chip it down to 25. I try to do a page or two of other IDs as well and I’ve only got limited time at the moment. It’s driving me crazy just how many observations don’t have decent enough photos for an accurate ID. How is it possible to get so much motion blur on an immobile plant? 
I guess quite a few are taken with add-on lenses on phones because they have some weird tunnel vision effect. Some data quality prompts would really benefit the upload part of this website IMO.
5 Likes