Recently i have been working on making a quiz, that uses iNaturalist API to show pictures of any species available on the iNat database. In my opinion, comparing to many quizzes i’ve played before, it has a few strengths that make it interesting :
It allows picking any taxa you want (and excluding some if needed)
Any location can be chosen for the sampling of species on iNat (based on an existing place or a drawn area on a map)
Pictures are selected (almost) randomly in all iNat observation of given species (only in the photo licenses that allow it), which avoids seeing the same pictures too regularly.
Has many other features that allow to adjust freely different parameters.
It does have a few downsides :
The max number of species possible in a quiz is 1500 (which i think is enough)
The loading at the start of the quiz, especially when working with big numbers of species, can be quite slow (up to 15 seconds from what i’ve seen, but usually 2-5 seconds).
To avoid sending too many API requests, i’ve added cache, so if you do a the same quiz many times you might start to see the same pictures come around at some point.
Other than training species identification and just having fun, i had a few objectives in mind developping this, which i think could be beneficial to iNaturalist :
Pictures are all from Research grade observations, which stop appearing in the “Identify” page. When you answered a question, you can directly access each observation by clicking their picture, which allows to quickly correct misidentified Research grade observations.
When selecting a place from the search bar (because this won’t work with drawn areas), you can access said place checklist after each question, which allows to correct mistakes of species status in that place. For example, exotic species shouldn’t appear by default in the quiz, but if they do you can then access the place’s checklist to add the “exotic status” so that the species won’t appear again in the quiz, and the status is corrected on iNaturalist.
Disclaimers : I don’t have the coding abilities required to create such a website and so i used AI in the making of it. I think i’ve respected all issues about photo licenses and API requests guidelines, but i might be wrong, and if so please tell me and i’ll try to address that fast. Also i don’t know if calling it iNat Quiz is fine regarding iNaturalist policy, but considering a website named iNat Guessr exists i thought it might be okay. i’d be glad to use the “Powered by iNaturalist” logo if i’m allowed.
The website is available at this link, i might get a domain later if people show interest in the site : https://inatquiz.pages.dev/
Please try it out and tell me what you think, and what should be improved. Screenshots of the site :
Cool! I’ve been working on my own quiz since the beginning of the year as well: https://www.naturvonabisz.de/fg-quiz/index.html. Initially, it was just meant for my university students so they could practice identifying species from real, variable field photos instead of always relying on perfect textbook images for their exams. It was so much fun that I added an Arcade mode, allowing up to 4 players to compete against each other.
Under the hood, my setup works a bit differently. Instead of fetching photos directly and completely freely via the iNat API—which often led to unsuitable or incorrect images for me—I placed my own custom database in between. This means not just any taxon can be played, but only those explicitly entered in my DB. This approach offers a few key advantages for my implementation:
Taxonomic control: I can determine via code whether an ID should be exact (e.g., no domestic pigs when asking for a wild boar) or if it’s allowed to include images of subspecies.
Filter management: I can set exactly what gets pulled per taxon. For butterflies, I forbid caterpillars; for amphibians, dead animals. Maybe that’s an idea for your quiz too! ;)
Image curation: The script prioritizes images from specific iNat photographers (favorites) to ensure high image quality before filling the rest up with random pictures.
Info database: The system allows me to attach additional metadata. After solving a question, it shows a cladogram for the species, plus info on its status (e.g., autochthonous, neobiota, domesticated), and soon, lifestyle and identification guides.
Exclusive content: I can password-protect individual topic areas—for instance, to test them or restrict access to specific groups (e.g., students of a specific course; accordingly, the University mode is password-protected).
Manual topics: Besides randomly generated questions, I can create manual question files that pull very specific taxa or ask special questions (e.g., “How many pairs of legs do brush-footed butterflies stand on?”). Thanks to the database, I can also control which solution is displayed and how, and of course, allow additional common synonyms that aren’t registered in iNat.
Before playing, the user can adjust several settings:
Focus & Difficulty (currently being reworked): You can choose whether to get only species, only higher groups, or a mix of both. There are 4 difficulty levels defining how easy a species/group is to identify from a photo. Thanks to the database, a taxon’s difficulty can be adjusted at any time.
Species filters: Domesticated, introduced, and extinct species can be toggled on or off with a click.
Intelligent distractors: In single-choice questions, the wrong answers are chosen to be as closely related to the correct solution as possible (the script climbs the phylogenetic tree for this). If a taxon lacks a common name, the question only appears if the scientific mode is selected.
Fault-tolerant text input: For text answers, special characters, capitalization, and spaces are ignored. So “Silver-washed Fritillary” is just as correct as “sILver – washed FRITillary”. You can also adjust how strictly the input is graded (e.g., strictly scientific, or including common names and an “easy mode” for kids—where “Duck” is accepted instead of “Mallard,” but yields fewer points).
Dynamic scoring: The system awards points differently based on the question type, chosen difficulty, and the given answer. Second attempts are possible but come with a point deduction.
I always show just one photo per question, but you can swap it via a button if needed. Besides single-taxon questions, the system also generates plural group questions (e.g., “Which family do these species belong to?”), loading a pool of images from various species within that group that you can click through.
Two more things I really care about: The entire quiz is fully responsive (works great on smartphones (I hope so xP)) and it runs purely locally in the user’s browser. There is no server-side caching and no user data is being collected or tracked.
Currently planned are a level system and certificates—and of course, more taxa and question types. If anyone wants to test my version: https://www.naturvonabisz.de/fg-quiz/index.html - use “Naturforscher” or “Artenkenner” difficulty for testing (and it is in German only atm) …and @tano_caperna: Maybe you’ll spot a feature or two in my setup that could be useful for your quiz as well! ;)
Very nicely done!
I like that it uses multiple photos from multiple observations, credits those photos, and allows excluding taxa, as well as tailoring the location. The user interface is clear and easy to use.
Your version, although having a slightly different purpose, is very complete (a lot more than mine !) and i will probably use some of your ideas in the future ! I really wanted to make the whole point of the quiz to be as exhaustive as possible, making it possible to train on any taxa, but it sure has limitations. I tried to avoid the issue of impossible pictures to identify by giving 10 each times so at least a few should make it possible to find the right answer (except for the extreme mode which only has 1 picture). Actually i thought before of making a category specially for larvae or particular life cycle forms, to learn specifically on them. I’ll have to see how i add that but i think it is possible directly via the API requests.
Making difficulty based on the taxonomic rank you have to guess is a really good idea and i think i will add that as soon as possible !
@tano_caperna and @fgallhuber I really enjoyed your quizzes! I am a beginner at identifying and I think both quizzes will help me learn! I tested them both out on mammals (figuring that might be the easiest for me), but I’m hoping to do bees and maybe other insects in the future. There were certain features I liked about each, so I hope to use both alternately! Thank you!
J’aime vraiment le fait qu’il puisse être limité uniquement à nos propres observations, ce qui, étonnamment, n’a pas été aussi facile que je le pensais . C’est un excellent moyen de se ressourcer et d’apporter plus de valeur à nos propres observations.
Avec plaisir ! Pour pimenter les choses, j’ai fait en sorte qu’on puisse limiter à nos espèces observées, mais les photos sont sélectionnées parmi toutes celles disponibles, de sorte que l’on ne reconnaît pas une espèce à la photo ;)
Very nice, this is what I’ve been hoping for! However, I do have a suggestion: what about allowing any of the known names, regardless of the language? Or, if you don’t like the idea, a possibility to choose the language for the species out of all the languages supported. I don’t mind the UI being in English, but my knowledge of species names in English is somewhat limited, and learning them isn’t really my top priority.
Cool idea! I played just a bit and noticed that in my special case it doesn’t really work, because most species I ID (neotropical aroids) are very context-dependent. So location’*, elevation, etc are really important. I guess you could have a small map with observation spread for a given taxon. But for a more general audience this would probably be overkill :)
‘* by which I mean something like “the pacific slope of Central America”
Hi ! Picking an area does not restrict the picture pool to the said area ! Only the species pool is extracted from the area, and then the pictures are randomly selected from all over the species range. Glad you enjoyed it though :)
. C’est un excellent moyen de se ressourcer et d’apporter plus de valeur à nos propres observations.