I don’t think the majority of users have an issue with some stuff. The issue is the balance and ratio of observations between the pine planted in a park and the cactus in a pot in the windowsill is problematic.
I honestly feel your pain.
The “clownish” nature of the data isn’t referring to the idea that researchers only study cultivated or wild things (the data of cultivated species can be used too, though it’s not as useful for as many studies) but to the fact that these two must be treated separately in order to be useful for researchers who use this data. If it’s a jumbled mix of the two without any way of distinguishing them, it’s either useless or requires so much data cleaning as to make it not worth using. This is why this topic is such an important one.
9 Likes
your pesky Pelargoniums are our indigenous ;~)
but your agaves and cordylines, and little potted succulents - Fred Ives?
I will ID my way, as best I can, thru wild and planted (in Cape Town), and it is a huge advantage that iNat is international.
2 Likes
Yes, it is looking so with no consistence, and we lack specialists because of that, it doesn’t affect new users who post 3 lillies from their garden, but it affect old users who need more special kind of info or just somethig more than agreement on a regular species. Cultivated should be marked, there’re many reasons for that and it doesn’t prevent studying it.
1 Like
I appreciate that @nathantaylor thanks for the clarification. I said it above, and I absolutely understand the need to mark things as cultivated. I’ve marked thousands (ok maybe just hundreds) of observations of plants as cultivated in Los Angeles. I just think it’s a little harsh to say the data is “clownish”. That reflects a bias towards biographic research (in my humble opinion) and doesn’t reflect the considerable onus on researchers using the data to vet and interpret the data they use in their analyses. Any researcher actually using other peoples data needs to be extremely careful with how they interpret these data–especially community science data. At least iNat has an open process for identification that can correct errors–a process which many traditional herbaria don’t have, or haven’t had until very recently. What seems to really be at issue here is that most folks commenting have a strong aesthetic preference for puzzling over the identities of wild native plants of local significance. And they would prefer to spend their time on interesting identifications rather than mindlessly clicking on “captive/cultivated”. I totally understand! My contention is that this may be a content moderation/user experience issue more than a data quality issue per se. I don’t see how it’s different from my experience trying to identify thousands of shaky cell phone photos of distant, nearly unidentifiable birds. Nevertheless those are eligible for research grade, but an absolute perfect photo of a rare plant in restoration site, with wonderful attention to anatomical detail isn’t (?). If we change our perspective ever so slightly, we could look at this as an opportunity to learn about and teach about cultivated plants–if we all did so, we would lessen the burden on other identifiers and spread botanical knowledge. I’m hoping that folks can accept that these cultivated records aren’t necessarily worthless (unless you are only interested in biogeography, and even then they could have value for the biogeography of other groups of organisms, or for understanding how non-native species get introduced!). Maybe we can actually leverage them to improve everyones experience. That’s my hope. With utmost and complete respect and admiration to everyone here. Who I sincerely agree with on almost everything. Also I’d like some help identifying cultivated plants on the UCLA campus: https://www.inaturalist.org/projects/ucla-campus-biodiversity?tab=observations. ;-)
5 Likes
I think that’s an overstatement. If you are interested in how the plant functions as a host plant, or interested in how much of it there simply is around, then it might not actually matter. Consider the case of Tropical Milkweed. It would be nice to know more about how much there is so we could study whether it’s playing a role in Monarch butterfly populations. Also other planted milkweeds. https://xerces.org/blog/tropical-milkweed-a-no-grow
Look, the distinction here for plants is partly, but not completely scientific. Birders (at least in North America) also have all kinds of rules about what counts as “countable” for species checklists. The first several generations of escaped pet birds are not “countable”, even though these intermediate generations (second, third, fourth etc.) would all be considered “research grade” on iNaturalist. For birders, they only become countable if they reproduce and expand without direct human assistance for several generations (even though many actually depend on non-native cultivated trees–which depend on irrigation… yada yada).
A similar rule could be applied to plants. That rule would be more restrictive than our current definition of research grade. Finding a single individual or volunteer in a garden would not count as research grade–arguably it shouldn’t for people interested in self sustaining populations. It’s always a little bit arbitrary. Or as another example, in arid regions a distinction between, irrigated/not-irrigated, could be equally important to the cultivated not cultivated distinction we insist on. I’m just saying these are all important and valid, but no one of them is universally more important. I apologize if I’m becoming unbearably annoying.
2 Likes
I don’t really disagree with you and “clownish” probably is a bit harsh, but I think it stresses the importance of this problem and how much the quality of the data hinges on identifiers doing the frustrating work of marking cultivated plants as cultivated. I’m hoping the developers find a way to address this. I consider this issue and the problem of school groups the most challenging issues to good data quality facing iNaturalist.
I have literally marked thousands too (just over 1,000 in CA alone). ;-) Looks like you can safely say thousands.
While I appreciate the sentiment, it’s a lot to ask of folks who specialize primarily on the native flora in an area and particularly of those who just use iNaturalist as a way to unwind. You might not think it’s too bad, but for someone like me who doesn’t know the cultivated plants and isn’t interested in them, it’s a major hassle. I also don’t think it’s fair to put a greater burden on identifiers who are volunteering their time and expertise. This is especially true if there are ways of improving the process so identifiers don’t have to take on this extra burden. Also think of it this way, every minute spent dealing with cultivated plants is a minute that could be spent in that experts area of expertise spreading detailed knowledge to the community.
Cultivated observations aren’t worthless, but they are a different type of data that are generally less informative about the things the iNaturalist community is interested. Keep in mind that iNaturalist is primarily for getting people to experience the natural world and not their own gardens. As for exceptional photos of rare plants that should be labeled cultivated, iNaturalist doesn’t do the best job of capturing that but you could include those as photos on the taxon page or link embed them in a journal post for identification. Cultivated observations are findable, just not the focus.
I think the key here is might. It’s true that some studies can use the data without this distinction, but for most studies comparing more than a single species, it matters a lot. Also, the data by it’s very nature is biogeographic, even in your example. Some host data can be mined but iNaturalist isn’t really set up for this despite numerous discussions in the past for some standardized implementation of this kind of info. Biogeographic data is the primary dataset that iNaturalist curates. Also, even in your example, having the clarity of wild/cultivated in the data would be valuable depending on what specific questions you want to answer.
This is true, and I certainly wouldn’t be opposed to a rule like this. The trouble is that it’s essentially impossible to confirm given the information supplied on iNaturalist. As such, I think the single generation will have to do. There has been a lot of discussion in the past about this topic and I think it would be better in a separate topic.
5 Likes
Sounds to me like changing the “Cultivated” flag so it doesn’t affect the “Needs ID” is the best answer. That way flagging isn’t perceived as penalizing the observation. You can flag observations all day and be viewed as only adding to the data rather than relegating observations. People that don’t want cultivated for projects or their ID effort, can filter them out. People that want IDs of cultivated plants have a better chance of getting it.
Is a cultivated observation “Reseach grade”? Depends on the research. Having the flag is the important thing. Knowing about the latest trend in cultivated plants may tip off a new invasive species threat. GBIF should filter them out.
Looking for the win-win.
9 Likes
Or, for that matter, a commonly planted native species! In my area, it is quite common for greenways or parks that are going for a more natural effect to plant native coast live oak, with the goal of eventually creating an oak-savanna habitat like the one that was there long ago. Does an oak titmouse care whether the oak tree it is foraging in grew by itself or was planted? Probably not. And this planting/cultivation of native oaks means that oak-dependent native wildlife will be in places it would not otherwise have been. So the “casual” (because planted) oak tree is then the direct cause of the research grade oak titmouse.
Bill McKibben’s prophecy is coming true: the distinction between nature and nurture is disappearing, as the continued existence of ecosystems becomes increasingly dependent on human actions.
I am one of those when it comes to identifying. And I honestly don’t mind IDing cultivated plants – although I will also mark them as cultivated after doing so. I love cultivated plants, too, because I know that I will never be able to travel enough to see all of their natural habitats; cultivated plants keep me connected to more of the world’s diversity than wild plants alone could do.
I always find plenty of the natural world in my own garden. Here is some perspective: a hobby apiculturist I know lives in a residential, suburban neighborhood. His bees gather nectar from whatever ornamental plants have been planted around the neighborhood. Their honey is classified as “wildflower” honey, because it is a mixture of different nectars, not from a single crop like clover or fireweed.
7 Likes
And that’s totally fine, just as long as it is acknowledged that there are people on here that do mind and would like the problem to be addressed. :-) In all honesty, a lot of the comments on here feel dismissive and propose the solution of simply being more supportive toward new users. If you want to focus on that, more power to you! We need more people like you! But for me, that’s not exactly why I’m here.
4 Likes
If iNat would encourage us to say - it is not wild, without removing it from needs ID.
And make it simple and effective for us to choose not to see not wild
It would work better for everybody
5 Likes
Or the reverse. Trying to figure out how to manipulate the filters to see casual observations is not exactly intuitive. I mean, I think I have it figured out, finally, but that’s after how many months of doing IDs?
3 Likes
Nathan thanks for the measured response. I am extremely appreciative of all your efforts and truly value your perspective on this. I was worried I’d provoke some less charitable comments. I apologize preemptively for extending the thread.
To cut to the heart of the matter, what I’m hearing is an identifier experience issue. We need identifiers to enjoy what they are doing. If you see a lot of boring or poorly documented observations, you will become bored/frustrated and spend less time doing the valuable work of annotating all these observations. I completely understand!
It only becomes a data quality issue if the observations are not being annotated correctly. We don’t expect observers to get ID’s correct, so I’m not surprised that they are not getting the cultivated/not cultivated distinction correct all the time. It’s not the end of the world from a data quality perspective, the system is built not just to sort taxonomically, but to vet other aspects of data quality (geolocation, date… & cultivated/captive). If a batch of observations being used for research includes some without the correct cultivated/captive annotation, I would be expect that even more of them are misidentified as well! In other words, a system that can’t keep up with a simple binary annotations is unlikely to be getting all the ID’s correct.
So I was thinking about this more (and apologies if someone already thought of this) but what if iNat rewarded/incentivized users for reviewing and annotating observations–just as it recognizes users for identifying observations. The positive feedback people get on social media (this included) is a huge driver of engagement (even if we can’t bring ourselves to admit it!). So we could create leaderboards and badges for “data quality reviewers” who vet and add annotations to observations. There are ton of people on this site who don’t feel confident about making identifications, so they don’t actually review observations much. If we told them that there are other equally important ways to improve the data, then we could get them more engaged and we could get more people checking observations. They would become an important part of the whole team. “Don’t know what it is? You can still help by reviewing data quality and annotating observations! Here’s how:” and then links to information on how to become a “super data reviewer!” I’m imagining, the site could simple count how often you click “thumbs up/thumbs down” on items in the “research grade qualification”. This could also be a good way to balance attracting new observers vs. attracting new reviewers etc.
One of the issues with the site as it is, is that it prioritizes reviewing only newly added observations which “Needs ID”. But observations can be corrected at any stage–even old observations with two or three agreeing IDs–need to be reviewed and edited when they are wrong. We need to find ways to encourage this kind of behavior. For instance, if you add an obvious ID to an observation that hundreds of other people could have also identified, that’s great we love you kudos, gold star, but if you correct an identification that two or three people got wrong–Wow, that’s someone who iNat should really want reviewing observations on the site. In other words, deemphasize ID quantity and find ways to recognize quality or thoroughness as well.
2 Likes
I’m sure no one wants me to extend this discussion here, but I just have to get in the last word–sorry :-) (I’ll stop after this). I still think you are missing my point. If I had a map of thousands of trees and shrubs in an urban area, with identifications, that would be extremely valuable for habitat modeling for a range of other species. It’s not about mining host plant data from observations of an insect, it’s about an ecologists downloading plant observations so they can better understand what’s going on in an urban environment, that data could inform studies of birds, insects, fungi, public health, even economics. I would argue that’s actually the strength of the iNaturalist data, collecting a huge number of general observations from all over the world (mostly in urban environments) and having them added continuously.
4 Likes
Well, that’s actually not where I’m coming from on this. It’s definitely frustrating but I see that more as a symptom and tend to think of this as an observer training problem. There are a lot of ways you could walk people through the process in a fun and educational way. That said, addressing the symptom directly with a leaderboard might not be a bad idea (sort of a “quick fix” perhaps). It wouldn’t really motivate me, but I’m a weirdo who tends to stay off the more common social media outlets like Facebook more than most people.
I guess here’s my question, do you think it’s important to mark observations as cultivated when they are and do you think that is useful for the kinds of research that iNaturalist caters to? I think you said that you did above so, if that’s the case, I don’t know why you are arguing with me here. My position is fundamentally that it’s important to mark cultivated plants as cultivated so that the data is more usable (and yes, less “clownish” :-) ). Let me be clear, I understand that there are applications of the data that doesn’t require this, but there are a lot that do, and I think it’s important enough that no one on iNat wants to compromise on that aspect of the data (perhaps not even you?). Can we agree on that?
3 Likes
Just as an aside, unless it’s a school group, three agreeing IDs generally is either very safely correct or more than someone without specialized knowledge can address. I would generally recommend more cleaning with two confirming IDs on RG observations, though for identifications, specialized knowledge is generally required for the majority here too for much refinement to be made. neither are not nearly as bad as the Needs ID observations, though.
1 Like
As someone pointed out long ago, it also skews the distribution maps.
Is that an ‘innocent’ horticultural industry garden plant?
Or a keep an eye on it garden escapee?
Or declared invasive?
Or wrongly and tentatively IDed (bit disconcerted that a single ID will still show up on the distribution map - encouraging a domino effect from Seen Nearby. Not. In fact)
5 Likes
Yes I agree. Mark them as cultivated! Please! I’m not disputing that. The point of iNat is to generate accurate data. If ANY of the fields are incorrect, then that’s bad.
That’s a responsibility that identifiers (especially experienced ones) share with observers. When cultivated observations pass through multiple reviewers and no one is marking them as cultivated (or think to ask whether it’s cultivated) that’s bad. I would imagine that this issue is similar in magnitude to misidentifications. But I don’t know.
So here’s what I propose, first I absolutely must stop myself from commenting on this thread any further (! sorry), second we need to actually dig into the data and find out how much of a problem this is. How frequently are research grade observations misclassified as wild? How does this error compare to misidentifications of taxa, etc.? It’s obviously going to depend greatly on the taxa involved.
My other point was that accurate records of cultivated plants have a lot of potential research value. Since iNaturalist is generating so many of these, and because this data is so unique, I think there’s an opportunity to embrace this as a feature not a bug. iNaturalist observations can be different from traditional biodiversity collections, and there’s nothing wrong with that. But I guess I’m doing a bad job here convincing people on this particular topic.
6 Likes