Hawk Talk: From Zooniverse to Here 

November 21, 2019

Big Red feeds a small morsel to the hawk nestlings.

Red-tailed Hawks are found across North America, and their calls are so iconic that moviegoers around the globe have heard hawk calls used by producers to depict Bald Eagles in films. Even though the raspy call of the Red-tailed Hawk and other vocalizations have been documented by researchers on the ground, Red-tailed Hawk vocalizations have not been well-documented in the nest itself. When researchers have checked nests in the past, they could only do so for a brief time to minimize disturbance. The 24/7 Cornell Lab of Ornithology Red-tailed Hawks cam gives us the perfect opportunity to document vocalizations right at the nest without disturbing the birds! 

Cam viewers and scientists recognized this after observing the nest in 2018 and brainstorming a scientific question to investigate. The community discussed several questions, but one rose to the top after a round of voting: “Do hawks use different kinds of calls in different situations at the nest?” With the question in hand, we set to work uploading 10-second clips into Zooniverse so that we could, as a community, collect data on vocalizations and things that happen at the nest. Zooniverse is a free and easy-to-use platform that hosts hundreds of citizen-science projects. On Zooniverse, citizen scientists watched clips and recorded both what they heard and saw at the Red-tailed Hawk nest during the first week after the first nestling hatched. 

Over 1,400 contributors joined the Hawk Talk project and completed over 48,000 classifications for over 8,000 clips. Each clip was classified by 6 people before being completed. After the community finished collecting the data, the research team set to work to extract the data from Zooniverse and decipher what the observations meant. 

The more people that watch a clip, the more accurate the data become, but it can be a challenge to work with multiple answers. With the help of Zooniverse volunteer and programmer Peter Mason, we figured out how to distill multiple classifications for each clip down to one observation. Since each clip was looked at by 6 participants and they didn’t always agree with what they saw, we needed to figure out a way to determine what actually happened. For example, let’s say that 6 people watched a clip and 5 people heard the nestlings “peep” but one person missed it. So 5/6, or 83%, heard a nestling “peep” and 1/6, or 17%, did not hear a “peep.” For each possible answer that participants collected data for, we calculated these proportions, which are called “vote fractions.” 

Once we calculated these vote fractions, we employed a 60% threshold to determine what actually happened in each clip. For example, if 4 out of 6 people (67%) classified that a clip had a nestling “whistle,” then we consider the clip to have a nestling “whistle.” If instead only 3 out of 6 people (50%) classified that a clip had a nestling “whistle” and the other 3 participants classified that clip as not having a “whistle,” then we are not sure if there was a nestling “whistle.” 

As we calculated these vote fractions, we realized that sometimes it was easier for people to identify that there was a vocalization, but not the type of vocalization. In addition to determining if a particular type of vocalization occurred for adults and nestlings, we noted if people at least agreed that there was some type of vocalization that occurred. For instance, if 3 out of 6 people (50%) classified that a clip had a nestling “whistle” and the other 3 people classified a nestling “peep,” we wouldn’t have agreement on the type of vocalization but we would have 100% agreement that there was some sort of nestling vocalization. 

After translating the collected data using this approach, we were left with 8,044 clips that we used to create exploratory data visualizations. It’s crucial to visually explore the data and look for patterns before we run any statistical analyses. 

Interested in learning more about what these data mean? Join us in our data exploration! Take a look at the data, explore different graphs, and share your observations with the community. Your observations and questions are key to guiding analyses and future topics to focus on. 

Check out this post that walks you through one of the graphs or jump right into the interactive graphs themselves.