Gameifying Citizen Science: Two Fun Ways to Get Involved
Do you want to get involved in science in a way that’s fun and easy? Do you like video games or platforms where you can track your stats and join leaderboards? Whether you’re a fan of games that get you out and active, like Pokémon GO, or prefer soothing mobile games you can do at home, there’s a citizen science option for you!
Today, I want to introduce you to two interactive apps that contribute to real, scientific research: iNaturalist and FathomVerse.
iNaturalist:
If you enjoy games like Pokémon GO that get you out of the house and interacting with the world around you, then you should check out iNaturalist!
iNaturalist is a nonprofit organization aimed at creating a broad data network for scientists and conservationists using individual observations. Using their built-in camera feature in the app, you can take photographs of animals or plants you observe in nature. Then, iNaturalist will suggest a likely identification based on the image and location data. Other users can then verify your ID or add different suggestions. After a classification reaches two-thirds consensus from users, the observation becomes “Research Grade”. These Research Grade observations can then be uploaded to databases for scientific usage.
“iNaturalist is an online social network of people sharing biodiversity information to help each other learn about nature. It’s also a crowdsourced species identification system and an organism occurrence recording tool. You can use it to record your own observations, get help with identifications, collaborate with others to collect this kind of information for a common purpose, or access the observational data collected by iNaturalist users.”
How to Use iNaturalist:
- Download the app and create an account.
- Snap a photo of an organism or upload an older image.
- Add a location and date/time observed.
- Check which species the app suggests and verify that it makes sense or change it if not.
- Upload the observation.
- Wait for other users to verify your identification or suggest alternate classifications.
- Achieve research-grade status!
iNaturalist also has plenty of stats and data to make it fun and interactive. You can keep track of your own observations and number of species observed. Moreover, iNaturalist also publishes leaderboard stats for different species, such as top observer and top identifier.
Additionally, they make observation data easily viewable on their website. You can check which seasons species are more frequently observed or to look at an interactive map of sightings.
iNaturalist Usage:
At the time of writing this post, iNaturalist data has been used in over 7200 scientific papers. Scientists (Hanson et al., 2026) found that including citizen science observations from platforms like iNaturalist is a cheap and effective way to improve the credibility of plans for establishing and expanding protected areas in the eyes of stakeholders.
CBS sat down with the app’s executive director, Scott Loarie, to discuss how the app contributes to science. Loarie also shared a fun story about when a user uploaded the first photos of a previously undocumented weasel species in the Andes. The very first photos ever captured of this species feature the weasel scuttling around on the user’s cabin toilet!
Loarie claims that iNaturalist data makes up a majority of the species observance data used now in science. Using the data uploaded to iNaturalist, researchers can track the spread of invasive species or monitor population declines or increases.
Logging data in iNaturalist can be done individually or in a social setting. Sometimes social events called a “BioBlitz” are held, where people join groups to log data in a collaborative setting. One BioBlitz event which focused on botanical gardens across Germany in 2025 generated 52,078 observations in one week.
GBIF and Scientific Application:
Research Grade observations captured in iNaturalist are shared with the Global Biodiversity Information Facility (GBIF). GBIF is an open-access data infrastructure which stores over 3 billion occurrence records and over 120,000 datasets.
Species occurrence data is frequently used for generating species distribution models.
Species distribution models (SDMs) combine occurrence data with environmental data to predict distribution across landscapes. Through the generation of SDMs, researchers are able to understand where a species is located. It also allows scientists to understand temporal patterns, such as migration or changes over time. SDMs require environmental data (like temperature, salinity, or phytoplankton presence in the marine environment) to predict where organisms might be found based on past observations.
SDMs are used in conservation biology and ecology. They allow scientists to better understand current circumstances or predict future changes in a species’ distribution.
FathomVerse:
Do you prefer chill mobile games you can play from the comfort of your own bed, or while waiting at the doctor’s office? If so, add FathomVerse to your library!
FathomVerse is a mobile game developed by the Monterey Bay Aquarium Research Institute (MBARI), the Monterey Bay Aquarium, and FathomNet.
FathomVerse allows users to engage with real ocean images collected by several scientific institutions. The aim of the game is to grow a library of labelled imagery which is used to train artificial intelligence (machine learning) models.
Images captured from my gameplay within the app. I’m “certified” in sponges, soft corals, and chimaeras! So if I get an image with those organisms, I can sort and classify them (hopefully) correctly. Left: the “image collection” process in the game. Right: the image identifier module, featuring a soft coral.
The usage of machine learning models is becoming more and more common in biological sciences. (Note: machine learning models differ from generative AI)
One large issue in science is that the body of existing data exceeds the manpower and funding available to analyze it. One research cruise can collect many hours of remote operated vehicle (ROV) video footage. Properly analyzing that data from start to finish takes a vast number of hours on top of the time required just to watch the footage.
One solution is to train machine learning neural networks to recognize and identify taxa. MBARI’s Ocean Vision AI is one example of these tools. The usage of machine learning tools can massively expedite the analysis process. Not only can this facilitate the analysis of new research, but also allow for data backlogs to finally be assessed as well.
Application of Machine Learning Algorithms:
Training AI models is an intensive endeavor which requires a large amount of input and feedback. The aim of FathomVerse is to engage a broader audience with game-based human annotation, while also providing education to citizen scientists who play the game.
Verified annotations and trained Ocean Vision AI algorithms developed from FathomVerse are published on open-source platforms like FathomNet. Providing broad access to these resources enables scientists globally to quickly and accurately analyze images and videos.
Visual data from the marine environment can be used to generate species occurrence data. This data can also be used in the previously-discussed ecological modelling methods.
Technologies like ROVs allow us to view and understand environments that are otherwise inaccessible to humans. And tools like machine learning facilitate the analysis of that valuable data, allowing scientists to better understand and advocate for these mysterious ecosystems. To learn more about the importance of the deep sea, check out my article on the importance of phytoplankton here (sounds counterintuitive, I know!).
Pin this post to read later!
The ROV we used on a research cruise in Greenland to gather video footage of an Arctic fjord. I used a portion of this data in my master’s thesis! We considered developing machine learning algorithms for taxonomic identification, but decided against it due to the scope of the data and topic. I ended up analyzing the images manually for percent cover, so the issue of time-consuming image analysis is something I am very familiar with.
