I have mainly Locals here in Maharashtra, India.
It is interesting that in Siberia we have several areas which colored as locals, but where almost all observations are made by “tourists”. For example Altai mountains. The most active observers travel to this site so regularly that they become more locals than in they home towns.
For anyone curious locals are purple and visitors are orange.
EDIT:
And looking at my area in Vietnam it’s clear that there are some issues with the data and the assessment of who is local vs tourist.
Maybe it’s better in other areas, but for my area I’d rate the data assessment at about a 7 out of 10. This is not surprising as this is something that’s difficult to properly asses, especially in a tourist area were some non-locals are repeat visitors and some people have a habit of posting other people’s content scraped from Facebook groups.
My area is fascinating. WAY more “local” observers than I expected. Also funny to note that you can pick out a lot of inat observers’ homes, at least when they make observations at home or in their neighborhoods. That huge purple blob looks to be a pretty prolific home observer. Also interesting to see that there appear to be other inatters in my neighborhood (and of course you can find my house, apparently through a few observations I’ve made where I forgot to obscure the coordinates). One person appears to be a local biology professor.
Hi all, this is Logan, the project creator. Thanks @radrat for sharing it here, and thank you all for the interesting observations and feedback.
As a result of this new interest, I am updating it to use data from August 2025 now. It might take a day or so, but I will post here when it has been updated, and you’ll be able to compare 2023 and 2025 data too.
The monikers “tourist” and “local” are just memorable ways to refer to different types of observations. Really, what’s being visualized is the duration of an observer’s history with the area around the observation (based on the h3 cells linked earlier). This is admittedly imperfect, as it must be, because the definition of a local or a tourist relies on extra cultural and personal context that’s not captured simply by iNaturalist observations.
An observer can have “local” observations in many places around the world, as long as they have a history of observing in each of them.
Some cases to think about/be aware of: an observer may be a world-class expert on arachnids of Orange County, FL, but if it’s their first observation, they will show up as a tourist (history in the area = one day). On the other hand, someone might be an enthusiastic amateur with no tropical expertise, but if they return to Costa Rica after taking a trip a year ago, their observations will now be considered “local” (history in the area = one year.)
Happy to answer any other questions!
Reading the blog, H3-3 tiles are used to group observations into “local” areas, which equate to roughly 87 miles in diameter. Since they are fixed hexagons, there could be some edge effect at work. If someone observed normally in one hexagon and travelled just a few km to an adjacent hexagon, their observation in that hex would be considered a “tourist” point.
Correct, that’s a natural side-effect of the aggregation method. However, as someone else pointed out above, if one or more additional visits to adjacent hexagons happen beyond a 90-day period, then observations for that user in that area will be color-coded as “local”, which seems like a reasonable cutoff.
After seeing this, I zoomed in on the north coast of the Dominican Republic. The purple cluster extending west from the word “Cambium” are mine; but I saw an even larger purple cluster east of there, in the Hoja Ancha Hacienda area, which turned out to be from a very prolific observer @mreith whose observations I see all the time when identifying.
Well, although I haven’t yet been able to make Cambium my permanent residence, the reason I keep going back is because I intend to be a local. I have a leasehold, and have submitted the drawings for a house. My avatar picture was taken there.
Okay, that went a bit faster than I feared. Thank you past me for leaving detailed notes.
The default map view now shows data from August 2025! You can toggle to view the older, January 2023 data for comparison.
Looks like there have been some new sketches in the Sonoran Desert:
This is fun, the display here might actually be the best I’ve seen so far for finding “iNatting hotspots”.* At very well iNatted places you can actually see where the trails are because the observations are so much denser there.
In southern Ontario there aren’t really tourist-dominated areas, Niagara Falls stands out a little which makes sense. I’m a bit surprised up much more touristed Montreal is than any Ontario cities. Pelee, Long Point, and Algonquin are all major naturalist destinations which essentially nobody lives at, but many naturalists will go to them every year which I guess makes them count as locals.
(*I’m thinking now that a really cool map would be one similar to this but that colours the observations on a scale based on how many species have been observed within e.g. a 1 km radius of the observation.)
Costa Rica (2023 - 2025)
Interesting that on the map, if you head West to get to NZ, the cells cut off at the international date line, so you have to make sure you go East to get to NZ.
Looking at it, the fact that Dunedin is all in one cell does mean that if I go just a few km South, or a little bit more West, then I would be making tourist observations. I rarely, if ever, go South of the cell’s Southern border. As you say, not perfect, but still an acceptable way of doing it.
Did you ever consider the idea of making it so that an observation is “local” if the person has other observations within some radius of that observation, or is that just utterly impractical?
I note with the H3 cells that if I go about 5km south and make observations, I’ll be a tourist, because that’s a bit beyond the city’s southern limit, and I really don’t go out there that much. Whereas if I made an observation and it was within a 5km radius of other, frequent, observations, I would be a local.
Amazing work, regardless :)
This is a lot of fun to play with! I also have speculated about the impact of tourist vs local observations in my area. It’s nice to see that in the data set.
Yellowstone observations are mostly by “tourists” - no surprise there. The hex with most of the park only includes a few smaller communities (West Yellowstone and Island Park). So “local” in Yellowstone likely means repeat visitor or worker. Many repeat visitors could live in nearby Bozeman, Jackson, and Cody — they would consider themselves locals.
You can easily see the main roads in Yellowstone, showing that most observations are made close to a road. There is more visible purple along the road in Lamar Valley - likely showing this is an area that attracts repeat visits.
There are some blobs at popular attractions and these generally have walking paths. This is the area around Old Faithful. The yellow seems more concentrated and purple more scattered.
In Grand Teton, you can see the main park trails. They even compete with the roads for most observations. I think this shows an interesting contrast with Yellowstone. Beyond the easy to access tourist trails like Jenny Lake, there is good tourist representation in the backcountry (Teton Crest Trail). However, the trails on the west side of the Tetons (in National Forest instead of National Park) are dominated by locals. This is just the southern section of Grand Teton National Park.
Overall, the model aligns with a lot “tourist” and “local” behaviors that make sense with what I know about the area.
This is probably pure fantasy, but I would love integration of with the iNat places to count tourists vs local observations for a given timeframe.
They mention in the blog post this was their original plan but it would be way too slow
I initially tried to calculate the number of observations by the same user within a radius of each of their observations. However, even with a Postgres/PostGIS query that seemed to be taking advantage of every indexing opportunity it had, this would take months.
I now see my own observations, and I’m a local. My two farthest apart observations are a Chinese privet (since destroyed) at a construction site in Shelby, and a spider on my windshield in the Westgate parking lot in Asheville, 102 km apart. The privet is present, but the spider is not.
I’ve quickly added a stats feature that shows observations over time. You have to be zoomed in quite far (far enough that the map shows every observation without clusters), but maybe this is interesting to folks too. As always, observations and feedback welcome!
You can see the seasonal pattern of observations in Portland, Maine:
The impact of corona on tourism in Quepos, CR:
And even the overall growth of tourism in Colombia:
