Big Bird Data
I’m a birdwatcher, and as such I’ve spent plenty of time outdoors with my neck craned upwards, straining through binoculars to figure out what the heck I’m looking at. This might seem far removed to the tech world, and thankfully most of the time it is. Mornings spent outside are a welcome reprieve from my screen-drive life, and I’d imagine the 45 million or so other birdwatchers in America (according to the U.S. Fish and Wildlife Service) feel similarly.
However, thanks to the efforts of the Cornell Lab of Ornithology and the National Audubon Society, technology is increasingly becoming a very welcome addition to the both amateur and professional ornithology. The Lab’s main effort, eBird, was launched in 2002 and has become a staple component of every birder’s arsenal. The site is an online database of user submitted bird sightings that provides users and researchers with live data on bird distribution and abundance. It’s become a shining example of citizen science, crowd sourcing avian enthusiast’s observations to help scientists track migrations, monitor threatened species and understand the consequences of climate change and natural disasters as they unfold in real time.
To understand how much of a “win-win” eBird is, it’s important to note that record-keeping is an integral aspect of bird watching. The primary goal, as much as there is one in birding, is to see as many species as possible, with extra weight being given to those that are deemed rare. As a ten year old, this scavenger hunt mentality really appealed to me, as I transitioned seamlessly from Pokémon to bird watching. Most serious birdwatchers keep a life list which is a list of every bird species they’ve ever seen, including details of when, where and how they first saw the bird. Many birdwatchers keep lists from every single outing they go on, and until eBird, maintaining all of these sightings was a haphazard affair. I myself transitioned from checkmarks in my field guide, to an excel spreadsheet, to a paid software called Birder’s Diary.
Then eBird came along, offering an extensive record-keeping service on par with any existing program, at the unbeatable cost of zero dollars. Cornell and the Audubon Society support the cost of the site through donations and government funding, allowing birdwatchers to use a service they would otherwise pay for in exchange for access to millions of observations from all over the world. Considering that the data in question is being used for conservation efforts, it’s a trade off that amateur ornithologists are happy to make! This aggregated data is also shared with the user, visualized in the form of range maps that show where a species has been found in the past and area-specific checklists that help to narrow down the likely birds in a country, state, county or even park. As of June 2018, 500 million sightings have been entered into eBird, and an estimated 100 million more are submitted every year.
One potential issue, quality control, is accounted for by an increased bar to the submission of rare bird sightings. If some prankster decides to report seeing a flock of flamingos in Central Park, then eBird automatically asks for photo evidence before accepting the submission. This precaution is extra important considering how many birders have come to rely on eBird for pursuing rare bird sightings. A false flag flamingo could send thousands of birdwatchers racing in the wrong direction!
Recently, Cornell has further harnessed eBird’s database with the Merlin Bird ID app. The app helps users identify birds, either through a checklist of characteristics (e.g. size, shape, location) or by submitting a photo of the bird in question. The app is powered by Visipedia, a computer vision project run by researchers at Cornell and Caltech, which uses machine learning to automate the classification of images. eBird’s database holds images for 95% of the Earth’s 10,000 bird species, and through user submission and reinforced learning its database is improving every day. Visipedia’s technology also supports the popular application iNaturalist, a database that aims to cover all of Earth’s flora and fauna, not just the avian members.
Looking forward, I anticipate that this technology will increasingly lead to the gamification of nature watching as automated photo identification allows for the incorporation of “scavenger hunts” into wildlife viewing. While these competitions come with their own potential pitfalls, I’m confident that any measures that grow the adoption of bird watching, and its accompanying appreciation of the natural world, is an overall net benefit to conservation efforts and society at large. As big data becomes ever more integral and imposing in our modern lives, it’s comforting to think that these same tools can be a force for preserving the natural world.
P.S. For those based in NYC, I’ve made the following Central Park bird guide for easy identification of the most common local species. Next time you’re on your daily, socially responsible walk outside, try and see how many species you can spot, and be sure to add it to eBird!
