Too Many Needs ID
Soon, there will be more Needs ID observations than Research Grade. This is unfortunate, as it is mainly 2 categories. We just do not have enough Insect and Plant Identifiers, so it would be excellent if Inat could begin an effort to make them Research Grade. Here is a breakdown, and you can see the problem. It would be success to bring the high numbers down even one percent.
42,940 Birds
20,813 Amphibians
26,312 Reptiles
24,459 Mammals
18,207 Fish
27,404 Mollusks
54,381 Arachnids
379,427 Insects
530,596 Plants
103,718 Fungi including Lichens
2,433 Protozoans (INCLUDES SLIME MOLDS)
90,157 Unknown
Thanks!
תגובות
Hey Danny,
This is indeed an issue... But it's the same in all collections, I think. All due respect to my backbone experts, but vertebrates are ridiculously simple to ID compared to the taxon giants like insects and plants... There are simply too many bugs and plants out there. I try to tackle the plants just in the state of TX, but there are over 5000 species! Keying out all of these plants from images is such a challenge, and there aren't really any good 'field guides' like there are for mammals or birds. Insects are the same way -- a stroll through the thousand of pages of bugguide show just how much of a challenge it is to give the 'right name' to an insect.
However, I think we may be putting too much credence on getting something to "Research Grade." In my opinion, it's totally ok if a beetle has to rest at family, or if a plant has to wait at a genus level. In the big scheme of things, that taxon level is sufficient for those wanting to document biodiversity or distribution of organisms. :)
Yes, I agree. Certain things are impossible to ID, but I think expansion and improvement of the Identotron and other features, and enabling a family to become Research Grade could reduce the problem of observations staying in the column unnecessarily . We could also make a quality grade for Not Identifiable. The needs ID column is way too long, and is filled with the observations you referred to.
Thank you for your insights!
DannyM
Fellow Naturalists,
If find this conversation interesting and hope you don't mind my butting in. As an old-school herbarium botanist (now enthusiastically embracing iNaturliast), it occurs to me that the Needs Id category is analogous to the indet folder in a herbairum. Most herbaria have them at the end of the genus or family for specimens that no-one could or did put a name on. Rarely is the specimen so poor that it just can't be identified. Mostly they are there because the collector did not know the plant and no one at the institution looked or knew enough to put a name on it. The important thing is that specimen was made and preserved. Many specialists or ambitious generalists (like my mentor Dr. Michael Nee) would go there first to have fun challenging their knowledge and skill. In many institutions, the indet folders contain undescribed species. I named Acalypha gentlei from specimens I found in the indet folders in the William and Lynda Steere Herbarium at the New York Botanical Garden (NY).
It's true that photographs, especially sub-par ones, are even more challenging than herbarium specimens. But I agree that if they can be realizably assigned to a genus or family, that's valuable data and should be graded accordingly.
One way to address the issue going forward is to help elevate the quality of observations. More and better identification guides will help. So will training. We could all do more to teach our fellows. Another idea is to designate stewards for particular areas. We are developing a program in New York City in which we will encourage people to adopt particular areas and help curate the observations from that area. Taxon curators are another way to work down the indets. Hopefully as more taxonomists come on board iNaturliast as i did, they will tackle the indets in particular groups. And there are technological "fixes" too, like the identotron feature. It would help if the "previous" and "next" buttons in the queries were tied to the most recent query, rather than the global query.
Great discussion. Thanks for letting me join in.
Daniel
Hey @danielatha ! Thanks so much for chiming in.
I too have some herbarium experience -- I worked for 3 years at BRIT (Botanical Research Institute of Texas) and even had the pleasure of going out collecting with Dr. Nee! He came down to work on a monograph of some genus in Cucurbitaceae.
I made a little journal entry about ID'ing stuff for others where I use my herbarium experience, especially with the 'indet' folders:
http://www.inaturalist.org/journal/sambiology/7871-the-nature-of-identifying-nature-appreciation-for-those-that-spend-time-id-ing-for-others
I wish wish wish I could get more of my academic friends to share their knowledge with the citizen scientists here on iNaturalist, but alas, there is unfortunately a lot of adversity. Fear that a digital collection will replace a physical specimen, perhaps?
I would LOVE to know what sort of techniques you are using to 'entice' your friends in academia there in New York to play along! I need some advice. :)
Also, there has been a LOT of improvement in the identify function here on iNat -- you can really play with the filters a lot to get a specific area and specific taxonomic grouping. @kueda @loarie and the other iNat legends have done a spectacular job in this.
~Sam
Also, @danielatha -- iNat still needs an observational record for Acalypha gentlei! ;)
http://www.inaturalist.org/taxa/518110-Acalypha-gentlei
Thanks everyone, I think that perhaps we could create an Indet quality grade, for Needs ID really should be for ones that can be identified.
@dannym, I'm glad you're looking out for the Needs ID observations! Research grade used to include things with agreement at or finer than family level, but it was decided that usually that's not especially useful, so they restricted it to species UNLESS someone votes in the data quality that it can't be ID'd. That's definitely a tool that you should use when you're sure that no one else could possibly ever identify something further (too far away, too blurry are the most common reasons I use it). Also, you can see when they changed the definition of research grade on the stats page (www.inaturalist.org/stats) because there was a big drop in RG observations.
Like @danielatha mentioned, I'd love to see more regional groups officially take the lead with iNaturalist like Vermont and Texas have, and NZ and Mexico have internationally. Unofficial efforts are great too.
After reading your profile, I have to say that my 6th grade self would have LOVED iNaturalist and I'm a little jealous that you get to grow up with it instead of learning about it as an adult! Keep up the great observations and IDs on iNaturalist!
Indeed! I'm glad you are with us Danny. Keep up the good work.
Thanks @carrieseltzer for the tip on voting something "unidentifiable" so it can get research grade at the family or genus level. I will look into it. To take the conversation a little further.... There has always been debate in the herbarium community about whether a difficult to id specimen (observation) is better off laying in the indet folder (in this case as "needs id") or whether it's better to have a name on it that is close but could be incorrect. The advantage of the latter is that someone with good knowledge of the species will find that specimen (observation) and correct it. The advantage of the former is that the species folder does not contain dubious specimens. In the herbarium we use a qualifier such as "cf" or 'aff" to designate uncertainty of an identification to species level. The "cf" suggests to "confer with" that species. And "aff" suggests "affinity" with. I was not part of the programming discussions Carrie alludes to and pardon me if these are elementary observations to all, but I'm throwing them out there in case it's new to anyone. Anyway, it's always useful to come back to this topic from time to time. Thank you @dannym for bringing it up.
And for using iNaturalist regionally, my colleague @brianboom and I recently got some startup money to develop a pilot for the EcoFlora of New York City project we envision being a sort of continuous BioBlitz of the City with participation of diverse communities and built around plants and their central role in the ecosystem (which includes us!!). We're starting with the grounds of the NYBG, but hope to expand after the first year. iNaturalist would seem like the perfect fit and we are looking for the @dannym s of New York City to participate.
42K birds? I need to spend some time knocking that down. I just did like 20 from the back end. I imagine the older records are where the most help is needed. I get a lot of practice with terrible pictures from being a moderator on the Facebook Bird ID Group of the World.
@vermfly , that was recently cleared by an automatic removal, but the old ones are still there. 20K are left. Thanks for noticing!
Thanks @dannym. I've probably doubled my identified in the last two days.
@dannym what do you mean by an automatic removal?
Or did 500,000 observs. become RG in 1 day? About a week ago Needs ID quantities dropped.
Maybe checking www.inaturalist.org/stats will help see what happened?
Looks like a big chunk of observations were entered around the 3rd and then identified over the next week. I wonder if that accounts for your large drop in "needs IDs", @dannym .
I see what happened. I set my location default to United States and by default it filters the Identify column to US observations. There are still 43K Needs ID birds.
@danielatha , will taxon curators curate specific taxon, for example one user curating Herons, another for Stink Bugs, etc. ?
הוספת תגובה