Welcome to Science with Shrike! Today we discuss citizen science projects, what they are, and how you can participate. We’ll also examine how they are the web2 version of “decentralized science”.
What is Citizen Science?
Citizen science is a broad term used to describe getting hobbyists, enthusiasts and other “lay” (eg non-scientist) people involved in scientific data collection. Citizen science projects are often organized by an academic unit or other non-profit, which manages the data collection, data curation and other administrative details. Funding comes from a mix of government grants and non-profit fundraising.
Why Citizen Science?
Citizen science is a useful way to collect large amounts of data that would not be feasible for a research team. Suppose you want to measure where a particular species of bird, perhaps the Peregrine Falcon, lives throughout the United States, and figure out if the range is expanding or contracting.
It is not practical for one person (or even a research team) to visit every location every year and try to assemble a list of sightings to get an idea of where these falcons are. But. Convincing people who go out looking for birds as a hobby (i.e. birders) to report all their sightings into one common database now lets you collect data across the world without having to go into the field yourself.
Once you have this giant database, you can perform all sorts of analyses, including training AIs to recognize birds. So Citizen Science shines in areas where data gathering is labor intensive, but a ton of hobbyists and other people are willing to do the data collection for you.
How is it web2 Decentralized Science?
For those tuned into the Decentralized Science, citizen science is the old version of decentralized science. Large numbers of people contribute data, the data are quality controlled (often by volunteers), and then are often freely accessible to the public. The organization (“protocol” if it was web3) manages the infrastructure and keeps everything running, just without a token or smart contract.
What examples of Citizen Science are out there?
The examples Shrike is most familiar with are related to distributions of animals and plants. But there is a government website listing over 500 Citizen Science projects. We’ll cover two major projects related to data collection for birds, and all other sorts of organisms, along with what each project has done with those data, and how they quality control the data.
eBird
Ebird is run by the Cornell Lab of Ornithology, and is the go-to database now for bird distributions, species photos and sounds. Most local Citizen Science projects dealing with birds are powered by ebird. The premise behind ebird is simple: provide some basic “effort” information (when you were observing birds, duration of observing birds, where you were, how far you traveled), and count how many birds you saw (or use “X” if you don’t feel like counting). Media (photos, video and sound) can be uploaded through a partnership with the MaCaulay library. Unusual sightings for an area are determined and vetted by a volunteer team of reviewers. There is no ID help, though wrong photo/audio observations can be flagged by super-users (those submitting >365 checklists per year) and reviewers.
Ebird provides many tools for those interested in birds. For the casual user, it tracks the birds you’ve seen, where and when you’ve seen them, and allows searching of media you’ve uploaded associated with the birds. It also allows searching for bird sightings around the world, so that you can plan your trips and target specific species. There is also an app for entering species in the field and tracking things like distance traveled, and time spent birding.
Cornell Lab of Ornithology has used ebird and the media to develop interactive range maps of birds and to train AIs to identify birds. Merlin is a phone app that will ID birds by photos or sounds. Merlin uses range to narrow the list of possible birds, and then give a best fit. While Merlin is far from perfect, it does a reasonable job, and improves every year. Efforts to identify birds by nocturnal flight call are ongoing, but more challenging.
iNaturalist
iNaturalist is another nature-related app, and focuses on all aspects of life. In contrast to ebird, iNat is more collaborative. Users upload photos or audio of the organism in question, provide geographic coordinates to their preferred level of accuracy, and it must be confirmed by at least 2/3 of community input on the identification. iNat has an AI that will try to help users narrow down the species, but it does not rely on range in the same way ebird filters do. The AI will get you in the right ball park most of the time, but not always.
iNat lets you submit any form of life, and you get feedback, so this is a great way to learn your local plants, insects, mammals, etc. Some time with a reasonable microscope (and formaldehyde to fix them) can get you neat protists as well. Certain groups are more readily identified, as mammals, birds, and butterflies/dragonflies seem to get the fastest IDs. Groups that are not well studied or cared about (fungi, grasses, tiny insects, protozoa/bacteria) take longer to get confirmation on the ID. Call a field cricket an Eastern Field Cricket, and you might be lucky enough to trigger the field cricket expert into providing you with a 100 page pdf on the nuances of field cricket ID and assumed distribution in the US.
iNat also allows users to create local competitions and “BioBlitzes”, based on whatever rules are desired. Often the rules limit the geography and/or the species that will count towards the game, but those are set by the creator. Overall, iNat’s listing function (keeping track of what you saw) is clunkier than ebird, and doesn’t break it down by geography as nicely. But it will track how many observations you have reported for everything.
If you have small children who you want to get excited about nature, iNat is the way to go. The app lets you take a picture and posts it for you along with location, and will give you suggestions.
Since both of these apps link a user name to geographic locations and time, it is possible to figure out patterns and locations from the data. While it is harder to download all the user data (and the APIs feel like they were written in the Stone Age), it remains possible to get an idea of where people were (and if they’re traveling). Ebird delays posting reports by an hour, but patterns of behavior can still be predicted. A checklist at 8 am every day at location X? Wonder where they’ll be tomorrow. Call your home list “YARD” or “HOME”? Hope it’s not pinned on your current home. You know this person lives in Michigan, but you found a checklist from today in New York? Bet they’re not home right now. You can use pseudonymous names, but Shrike recommends against a pseudonym that you want to keep mostly anon.
There are many other citizen science projects out there, with different purposes. Get involved, and your hobby may contribute useful data to increasing our knowledge!