Observation density

[anneclewis]

I am in the process of putting together a series of lessons on using iNaturalist data for middle/high school geography educators. An objective of one of the lessons is what might be the geo-political context of making observations? Population density, development, access to internet, are all factors.

One of the questions I have is about what non English language users see when the app is used, particularly outside North America/UK. How much does language figure into observation density in other countries?

For example, do Chinese iNaturalists see the app interface in Chinese? It looks to me as if the IDs are in English but maybe not? That would make a working knowledge of English one of the factors.

One of the most interesting examples of observation density is South and North Korea. I think world geography teachers would be really interested in the factors that played out there. What are your thoughts?

[andy71]

I really like your idea of using this to teach about Geography. It’s a great idea because they can explore the map and see the patterns themselves.

Besides South Korea vs. North Korea there are tons of local examples as well were you will see large differences in observation density. Sometimes this is due to observer preference, sometimes this reflects lack of access and sometimes this reflects differences in biodiversity.

[cmcheatle]

The site is translated into approximately 30 languages. At the bottom right of each web page is a drop down where you can select the language to run.

The mobile app picks up the language the device is set to and defaults to that. I believe if that language is not one the site has a translation for, then I believe it defaults to English.

Different languages have differing volumes of translations of common names. Some are very complete, others are barely done.If a user enters an ID in English on an observation in China, assuming the species has a translation entered, and the Chinese user is running in Chinese, they will see the Chinese name. If no translation into Chinese is entered, they will see the scientific name.

So for example while living in Canada, I run the site in Danish. As you can see here, most records show me the scientific name, but the one in the bottom left which does have a Danish common name entered, it shows me that.

[pisum]

maps can really be beautiful and interesting. below are some screenshots of a map i put together quickly in ArcGIS Online. the first screenshot is just the standard AGOL topo basemap. the second is an Earth at Night layer from NASA’s Visible Earth, which can help visualize population centers (and wealth). the last is a GBIF occurrences layer on top of the Earth at Night layer. (i used GBIF instead of iNat because it should include iNat + other sources, and because i like the way GBIF’s density maps display better than any of iNat’s observation maps if you’re going to be looking worldwide.)

there are lots of interesting questions that can be asked just based on comparisons between these 3 screenshots. here were the first few that came to my mind:

1. why do those Scandinavian countries have relatively so many observations vs. Russia, just next door? (look at how clear that boundary is.) is there a linguistic break there? or some social break? or is it purely political?
2. look at Taiwan vs China. both are Chinese-speaking countries, but why are Taiwan’s observations so much more dense? i figured that if you just looked at iNat data, there would be a difference because China’s Great Firewall blocks Google Maps, which greatly hinders iNaturalist’s functionality in that country, but maybe the Great Firewall also blocks other apps, too? but maybe the tech factors really take a backseat to other factors like politics, education, economy?
3. looking at observations, you can sort of see where the US-Mexico border would be (whereas there is no such obvious break between Canada and the US). is that due to language or politics? or is it just that because there’s a giant desert there? (deserts in other parts of the world also seem to have a low density of observations.)
4. the dark spot in North Korea is interesting when compared with South Korea, but it’s not the only one. look at South Sudan and Turkmenistan. there might be others, too.

oh… also, you may want to check out the series of World Tour discussions and blog posts that loarie has been spearheading. people talk about how iNat is being used in different countries. sometimes you get interesting reminders of geopolitics, such as something related to Crimea here: https://www.inaturalist.org/blog/26633-ukraine-inaturalist-world-tour.

[apseregin]

GBIF data are biased. One should be within GBIF Secretariat to know all obvious and hidden circumstances. The main one is actually a national membership of a definite country in GBIF. For instance, Russia and China do not hold national GBIF-nodes, that is why you see virtually no data from these areas (but probably, one can see an orange spot nearby Moscow - it is a dataset for my thesis with 123K records). Please, consult https://www.gbif.org/the-gbif-network to see the correlation between membership and data density.
Turkmenistan is 95% a desert like Sahara.

[pisum]

@apseregin

GBIF data are biased.

i suppose any data source would be biased, but if you prefer to see just iNaturalist, below are the observations from iNaturalist sent to GBIF. (so this should be just a subset of Research Grade observations from iNat that are at least a certain age. i still didn’t get the maps directly from iNaturalist because i still prefer GBIF’s density maps.)

based on this latest view, i no longer really see the distinct Scandinavian hot spot, and i don’t see the distinct South Sudan and North Korea dark spots. instead, i see bright spots in North America, Western Europe, New Zealand, Taiwan, Hong Kong, South Africa, Australia, Israel, Singapore, South Korea, Russia, and a few other places.

to me, this looks a lot like a global wealth map, with maybe a slight bias in favor of countries with iNaturalist portals (the original US, Mexico, NZ, Israel, etc.). other than that, i do notice some places that are probably destination nature spots – various islands are particularly bright, Ecuador is bright, Tanzania is bright, the giant parks in northern Benin and neighboring countries are brighter than the giant West African cities along the coast, etc… so maybe it’s like a wealth map, mixed with a rich person’s vacation map for nature watching.

the distinct US-Mexico border is still here, too…

EDIT: in case anyone else wants to recreate this in ArcGIS Online, here’s the tile layer URL i used: https://api.gbif.org/v2/map/occurrence/density/{level}/{col}/{row}@1x.png?srs=EPSG:3857&style=purpleYellow.point&publishingOrg=28eb1a3f-1c15-4a95-931a-4af90ecb574d

[anneclewis]

Thank you to everyone for their thoughtful, useful contributions. I realize the data you get from iNat is coarse, really at the level of presence/absence so it doesn’t have a lot of applications in GIS analysis. It’s good to know that there are ways it can be taken further if a teacher or motivated student were interested in such things.I am bookmarking this thread as an exemplar for teachers who want to go further in analyzing iNat data.

[pisum]

i actually think that teaching kids how to create a map would be a really useful skill. with the wealth of available data out there and the easy-to-use tools available (like ArcGIS Online) nowadays, it really only takes a few minutes to create a map with a couple of datasets, and once you know how to do one, you can apply it to so many other purposes, whether environmental science, public health, social science, marketing, etc… if you’re in the USA, the elections are coming up, and there’s nothing more satisfying than creating an election map and throwing that up against some sociodemographic data to see how one affects the other.

if a given school has an animal mascot, or if there’s a state flower or a national bird, or something like that, it might be fun to find where such an animal, flower, or bird lives using iNaturalist data. maybe if you can find historical ranges, you can compare and talk with the students about how to get the ranges back to what they used to be and maybe even take action in that county? (maybe get the students in touch with a local conservation group or the county ag department or something like that and see what kinds of things they might be able to help with?)

anyway, good luck with your lesson plans!

[shelley_b]

Wealth does correspond strongly with internet access/affordability and camera ownership.