Though I’ve used iNaturalist sporadically for years, this was my first year being involved with the City Nature Challenge. My “City” skipped the CNC last year, and I ended up as one of the local organizers this year since it didn’t seem like any other organizations were going to step up for it. I had a lot of fun, and it’s motivated me to start using iNaturalist much more regularly.
It was disheartening, however, to log into the forums for the first time yesterday to make a bug report and see the lengthy threads on how the CNC and similar projects have negatively impacted dedicated members of the iNaturalist community and the data sets (see 1, 2, and 3). While I hadn’t personally seen many of the reported issues (falsified data, re-posting of copyrighted photos, etc.) it could simply be that I didn’t know how to recognize them.
Do you have suggestions on how I and other local organizers can help improve things? While there were suggestions to address these issues on the other threads, they were largely targeted towards iNaturalist itself or the global organizing team. I’d like to (1) pitch in on cleaning up my City’s dataset for this year in a more thorough way and (2) try and stave off as many of the problems for next year as possible. I intend to share any suggestions in this thread with other local CNC organizers as well.
First a gentle reminder - the Forum is where we (few!) come to whine and troubleshoot. Thank you for organising a CNC! I also came across few problems in Africa. Depends whether Your Slice of iNat is a good one or a dud.
Please identify. For your preferred taxon. Or your location. Or your taxon location. Apart from Unknowns, I was horrified when I first tripped over about the same volume at Needs ID.
Please encourage other taxon / location identifiers. Every little helps.
Ultimately - work towards clearing the ID backlog before the next CNC.
We have various tools that we can use. Among the low hanging fruit now are CNC obs at very broad IDs - which missed the taxon specialists who tend to filter by Family. Lots of plants and dicots - which are easy for local identifiers.
One improvement generally is that all or really most local organizers should be taking a leading role and responsibility for their projects. They should be on the front lines in observation cleanup afterwards.
This involves correcting observations that are obviously wrong like ones IDed as species that dont occur on the continent, ones that just dont seem right, ones that are too blurry but IDed to species, etc.
Pushing observations that are likely incorrect and have become RG due to blind agreement, attempting to prevent mass blind agreeing. Trying to audit RG observations especially if it seems many are becoming RG due to blind agreement.
Flagging copyrighted observations, marking observations captive, reporting inappropriate behavior from participants like possible sockpuppetry, mass copyright infringement, or other issues. Using DQA on observations where appropriate.
There are likely other things that can be donw, but this is some things that should be done while and after the CNC.
I thinks she’s trying to say that here at the Forum we’re far more likely to discuss problems than the normal good working of iNaturalist. It can give you a biased idea of what’s going on. For example, I suspect most CNC projects have very little deliberate falsification, though a few seem to have a lot. Which do we spend time writing about?
Are the copyright violations you are seeing fairly blatant (the same photo gets ripped from another iNat user on the same project in short order, watermarks that don’t match the username, etc.) or is it something where you might need to do a little more digging to discover it? I saw someone mention photos being stolen from eBird, but since I’m almost never on eBird I likely wouldn’t immediately recognize it as being a copyright violation.
Can you also clarify what you mean by “sockpuppetry”? Is it just people making multiple accounts to agree with their own IDs?
What can organizers do? Every single CNC should allow only Needs ID and Research Grade observations to count toward totals.
This won’t immediately stop the postings of cultivated plants, of course, but at least they’re gone as soon as they get marked correctly. And just scanning for garden plants will be a quick way to help trim the project down to size. (I’m sure some projects do this already, but perhaps not the biggest problem projects.)
National Geographic sort of photo (Okay once it turned out the the photographer is on iNat, it is his photo) I use right click on an image for Google Lens, but there are other tools.
Otherwise, while identifying - I have seen that photo before …
I agree with @sedgequeen. I happen to like using Casual observations for my own CNC observations, but I will gladly give them up to help out everyone else. It’s the only quick and easy solution I’ve seen that might make a difference.
Rosa - Rosa - Casual is a valid obs. With 2 sets of CNC numbers for Casual to encourage newbies learning to engage with nature. CNC organisers need identifiers for cultivated plants, captive animals.
And another set of numbers for RG according to iNat’s guidelines.
Where Rosa - Rosa - RG is a problem, unless it fits the guidelines.
Thanks for the clarification, and for linking to the specific tools! The geomodel is totally new to me, and it seems like a fair number of the other local organizers are new-ish to iNaturalist and may not know about those features.
Kids supporting the home team? Insisting that rose is a camellia, because my friend said so. Was not explained to them that camellia is factually wrong, for a rose.
Because the Community Taxon algorithm requires more than two thirds to be convinced
2 = RG
1 wrong = 3 for RG
2 wrong = 5 … that is when identifiers burn out.
Thanks for asking these questions and organizing your own CNC and caring about it!
I have seen all of the above. In my experience, most are images ripped off from other online sources. I suspect that they are just whatever sources a person finds most handy. Some are one that come up in image searches (like Google), sometimes a user will go on a binge from a source (Wikipedia, eBird, some local site, whatever). This year, we saw more of users seeming to intentionally share or copy large sets of photos from other CNC users wholesale (so like one user took photos, which would be copied and reuploaded by tens of other users). Some users will crop the photos weirdly (or crop out watermarks) so it isn’t as obvious that they are copyright infringing. Some new users may not know that they need to use their own images, but especially in cases where users are cropping or continue to upload infringing photos even after being warned, it seems very likely that they are doing something they know they should not.
Yes, creating a second (or more) account for any purpose (multiple accounts for a single user are not allowed on iNat except in restricted circumstances). Mostly these are used for adding IDs to one’s own observations (or friends’) but also sometimes for trolling/harassment. In the context of CNC, what many users have reported is friends/sockpuppets agreeing to computer vision suggested IDs that may be wildly wrong. Because of how iNat’s Community ID algorithm works (which I generally like!), overcoming bad faith IDs is particularly time/effort-intensive.
I think this is one of the overall issues, to be honest - many of the organizers don’t have a real feeling for how iNat works, and so they may not know how to guide their participants either.
I’ve messaged or tagged a large number of project organizers about issues during this CNC, and only a tiny handful have even responded. I suspect some of them don’t even know how to check their messages and mentions (maybe they’re only using the app, I hear it’s easy to miss those things on there).
But please don’t take the complaints about the overall CNC issues to be directed at you - the majority of projects this year have had very few issues at all. It is just a few projects, which unfortunately also happen to be very very large, which have overwhelmed us all.
I am working on a set of suggestions for how to mitigate some of the issues in the future - I’ll post it as a topic and invite others to contribute soon.
Having looked at your profile, I can confidently say, you are my home team and I have not seen any issues with our team. As others have said, there are many teams with no issues at all. I would not assume any team has issues, until I have seen them first hand.
Nevertheless I would recommend
Never pressure anyone to participate. By this I mostly mean, do not assign students to make observations, especially not for class credit. I also read a while ago that La Paz required their government employees to make a minimum number of observations per person, which would be a similar but less likely scenario. Once people feel required to participate, the subset of them who actually would rather not participate become inclined to cheat. On the forums these people are known as “duress users.” You can search that and see lengthy discussions about it.
If you want to take the data aspect of the challenge seriously, set your project to include only “needs ID” and “research grade” observations. This will exclude “casual” observations. Casual is a large category lumping cultivated plants, captive animals, observations missing data (such as no media), and observations with false data (incorrect location or date, copyrighted media, etc.) Some people find some value in some forms of casual observations, but there’s no way to count just those observations and not count the truly falsified. For example, some observers enjoy making no-media observations in good faith (“I am sure I saw this but could not get a picture.”) Some people (myself included) find some value in cultivated plants (while others can’t stand them.) But if you want any causal observations to count towards your totals, then they all will, even the ones completely falsified. Again, falsified observations are an issue for some teams but by no means every team.
It is good idea to find active users (most and foremost identifiers) and ask them to participate in curation. Or contact local ornithological/botanical societies - many are willing to help or even have already organised similar activities on different platforms. In my city most events by our local society are organised in social media, through e-mail or offline, but they are not particularly active on iNat.