this screenshot is from the NYT newsletter, California Today. It reports on a story by The SF Chronical citing the most “unusually common”* mammals found in Bay Area Urban areas by city. It quotes @damontighe about the use of iNat data. The San Francisco Chronical did the study using iNat Data. Since this seems to be paywalled, here is a screenshot.
I also managed to see the original story by Susie Nielsen and its data breakdown by poking around the Chronical website and searching for stories on her name…
Interesting story, though nice irony in using free, citizen science data, much of which is offered under open licenses, and then paywalling the results…
There ought to be a licensing clause where publications (scientific or otherwise) using iNat data have to be open-access. Many journals and newspapers alike now offer a mix of open and paywalled content, so it’s not like it would be twisting their arm to do something they don’t already do.
Although I’m not currently a subscriber to the SF Chronical, I was able to access the story. I clicked the Close icon on the subscribe pop-up, and I could read the story and see the data tables.
Reminder that if you don’t want your observations included in paywalled studies (for commercial gain), update your observations to CC-BY-NC in your iNaturalist account settings.
The unusual animals associated with particular cities in this article are often not observed in those cities at all, but rather in areas well outside the cities. For example, there are no Tule Elk in Petaluma, but if one selected an area around Petaluma that included locations near Pt. Reyes where obscured Tule Elk coordinates would show up, and didn’t really know how iNat works, one could be fooled into thinking there are. Similarly, there are no Elephant Seals in Inverness, but there are in a search box one might draw around it.
Also, I happen to know that the large number of opossum observations in Glen Ellen are attributable to a single opossum that likes to wander back and forth in front of a camera trap there.
And the association of Brown Rats with Oakland is driven by observations by a single observer on the running path around Lake Merritt. Almost all of these associations are easily explained away. Berkeley really does seem to have a lot of Fox Squirrels though.
For most taxa a small number of “superusers” account for a disproportionate amount of the observations, leading to significant bias if you are don’t take that into account.
Yes, an unfortunate stereotype. I’d say it is a bit irresponsible of the SF Chronicle to publish this “analysis” at all, even with somewhat buried caveats, given how poorly they understand the dataset and the obvious limitations of their approach. It is useful mostly as a cautionary example of how to not use iNat data.
Berkeley’s fox squirrels are the hundreds of observations made on the university’s campus, where the squirrels have their own social media accounts and have been the subjects of long-running research studies as well as many short term class projects.
How on earth are they considering northern elephant seals, Virginia opossum, and wild boar “unusual” animals for the region?!
All three are extremely common in the greater SF Bay region. The northern elephant seal is so common you can see them on certain beaches in nearly every Google Earth image update.
Wild boar, several types, are all over the place, with signs of them blatantly obvious.
And opossum are very common in backyards and in & under infrequently used sheds (and underneath a small house I was renting in the area a long while back).
How on earth are they considering northern elephant seals, Virginia opossum, and wild boar “unusual” animals for the region?!
Apologies. I think that was me badly expressing the gist of the story in my title, which I have since corrected.
We defined whether a species was unusually common by looking at its sightings as a share of all sightings of its taxon (i.e. mammals or birds) in each city, and how that compared to the share of sightings of that species in the Bay Area as a whole. For example, coyotes represent just under 1% of all mammal observations across the bay, but 3% of all San Francisco observations. We only included species that were sighted at least 40 times in the last three years.
Unfortunately, I don’t think that would help in cases like this one. They just performed searches on the explore page, without understanding any of the underlying observations. I very much doubt they filtered based on the licensing.
That’s, unfortunately, normal. Organisms that catch people’s attention will tend to be reported more often.
It’s the same sort of bias that leads people to think that it rains more often when they don’t bring an umbrella; they remember those times because they stand out, and forget all the times when they brought an umbrella and it rained even if the latter vastly outweighs the former.
