That would only work if there was some way to ensure that both instances don’t get flagged as duplicates.
5 Likes
We are managing to deal with the ‘single subject’ DQA. We also have to withdraw the DQA there if the original multiple subject obs is tidied up to just one survivor. And we have had that DQA for months now.
That’s good.
While I am glad we have the “single subject” DQA, it is not free of problems (people interpreting it as being meant for observations where there are multiple potential subjects, not different subjects in different media items). And because some users still look at casual observations, it does not really solve the issue of how observations with photos of different subjects should be ID’d.
But a “duplicate” DQA would be far more difficult, precisely because “duplicate” is not about the characteristics of a single observation, but how one observation relates to other observations. Because iNat’s database treats observations as independent entities, a “duplicate” DQA would have a lot of potential to be wrongly used – even with the best intentions. All of the following scenarios seem to me to be not only plausible, but likely:
- DQA’ing of all observations with the same photo, not just all but one of them (e.g., because two people are looking at the set of observations more-or-less simultaneously and choose different observations to DQA without realizing that someone else is also doing so)
- DQA’ing of observations that have been duplicated using the “duplicate” option or media that has been re-uploaded to focus on a different organism (this already causes occasional confusion when IDing, particularly if the organism of secondary interest is in the same broad group as the more prominent one)
- DQA’ing of observations of the same organism made by different people (as forums discussions indicate, some users already think these should not be uploaded)
And because DQAs do not generate notifications, any wrongly DQA’d observations would become casual without the observer having an opportunity to clarify.
The only way I can see any mechanism for handling duplicates working would be if the underlying structure of the database were radically restructured.
…and this strikes me as rather off-topic and not particularly relevant to CNC specifically, as most duplicates are the result of errors of new users and not people intentionally trying to manipulate the statistics. It is also something that would have to be tackled on iNat’s end and not by changes to the organization of the CNC.
3 Likes
This sounds generally good. More time to curate and ID will be great.
I’m less of a fan of excluding captive/cultivated, as a lot of nature in cities is captive/cultivated, and for plants a lot is unclear whether it’s wild or captive/cultivated. If people go into a city forest reserve that was planted up 50 years ago, none of the photos of the adult trees are going to count for the CNC, yet knowing what’s still there and what’s in flower and fruit is really useful knowledge for cities.
At least that’s my thinking when I make observations of planted native plants across my city during the CNC. I don’t regard these observations as of a lower data quality. One more observation of a wild house sparrow is a lot less valuable.
Having said that, I understand the motivation not to flood the CNC with potted house plants and pets.
4 Likes
I am sorting out duplicates for Great Southern Bioblitz. Now, till 10 November. The same issues with CNC.
mostly, people have now learnt what the DQA is intended for. Similar issues around ‘Not Recent’ DQA which is intended to make fossils Casual.
Applying the DQA is on the people who use it. My case use is not your examples. John has uploaded the same picture, at the same time, for the same taxon.
Most CNC observers will be in cities (it is in the name after all). For Newbies - whether what they are looking at is Wild in iNat’s jargon is an enigma. The obs will still be there, but, not counted for CNC - which discourages cheating, in intention. Maybe we could have a parallel Umbrella Project for the Not Wild obs. And elitist to dismiss the biomass of deliberately planted, if we intend to encourage people to engage with nature. That street tree sustains some biodiversity.
2 Likes
If people would like to continue a conversation about the Data Quality Assessment/duplicates/data quality issues, please start a new topic.
3 Likes
I agree on the value of some cultivated/captive observations in the situations that you’ve mentioned. I also agree that there’s value in interacting with nature via captive/cultivated organisms as in parks that many urban users will encounter and some of the data generated from those.
But that said, having seen the issues with past CNCs with a torrent of captive/cultivated observations that overwhelm IDers and represent the large majority of observations in some areas, I think there needs to be a change to reduce this behavior. There’s definitely a cost as you’ve highlighted, but I think the potential benefits outweigh that, and excluding captive/cultivated is something that at least needs to be trialed.
10 Likes
Realistically, about 2-3 months (or more) are probably needed to really handle IDs and sort out the bad/ineligible observations from the good. I could even imagine putting off a final decision on results until late in the year. But that’s probably not something the organizers are willing to do.
3 Likes
I believe one project has over the course of a few months jumped 10% causal (if not more) through DQAs, copyright flags, captive, etc. Time is needed to sort through observations. That’s really my main concern at the moment. The other changes are quite promising in my eyes.
4 Likes
Maybe we could have a parallel Umbrella Project for the Not Wild obs.
That’s an excellent idea. It would keep the CNC focus on wild things while still encouraging people out in urban habitats to observe all the species they find. I like it.
4 Likes
Thank you so much for showing your commitment to collaboration and community building.
4 Likes
I would support that, but how do we motivate people if it is not part of the Challenge?
It is alright for those regions where trees are part of the wild life. But in southern Africa, Durban has indigenous trees, but Joburg, Cape Town, Bloemfontein occur in grassveld and fynbos that has no trees, so the entire urban parkland and suburbs are excluded from the Challenge: there is NOTHING plantwise apart from a few pests and invasives. Yes, there are some animals and fungi living in or off them, but it effectively excludes city plants from the Challenge, except in a few nature reserves. And that includes plants planted to attract birds, butterflies and bees.
At the same time we are losing trees to the Polyphagous Shothole Borer Beetle, and monitoring the health of these trees (and the increasing area of their devastation) is really important. Except that these are planted Oaks, Maples, Poplars. We are having an “ecosystem” shift in our cities as trees vanish and unaffected trees remain. But we cannot record it for the City Nature Challenge!! We use the Great Southern Bioblitz in spring to monitor indigenous and threatened species when they are flowering and most obvious. The City Nature Challenge in autumn was therefor focussed on urban landscapes, which now are no longer part of the process.
There are other diseases of urban planted trees that will be worth recording too.
A great pity …
1 Like
You can upload anything you want, anytime. Casual records don’t typically count on leaderboards or show up by default in searches, so the new CNC rules better match the day to day user experience.
3 Likes
We could make a parallel (collection) project CNC26 Not Wild (but still minus cats, dogs and people). Then those projects can be compared year by year.
Still hoping that one day iNat will do - Wild - Not Wild - and Broken (missing data). Not Wild at Needs ID versus CID (not RG). Then observers can get IDs, and identifiers can avoid Broken and Not Wild. Status quo is small chaos. This is from today, to prove my point https://forum.inaturalist.org/t/can-a-reminder-be-circulated-to-members-about-need-to-mark-cutivated-captive-v-wild/72472