“I’ve heard these adorable stories of little kids seeing the sleeping animal pop up on their parents’ Fitbit or Google Pixel Watch and knowing that it’s bedtime,” Elena says. And the team will continue to learn more, like which animals tend to overlap the most (for example, giraffes are most likely to transition into dolphins).Īll of this work is part of their larger effort to take this extremely useful data and make it something people can act on. Tortoises (who tend to take longer to fall asleep but spend more time in bed) are most often women. The dolphin (which is associated with lighter, shorter sleep and is the most likely to have long periods of being awake) is the least common sleep animal.
As of September 2022, the giraffe (which tends to have shorter yet deeper sleep periods and early wake up times) was the most common sleep animal. They also dove deeper into the data to figure out how many and what kinds of sleep types would be best to help people learn about their sleep health.Ī year since Sleep Profile’s launch, the team is still learning about sleeper types. The team worked with ethnographers and animal experts to determine the right animals for each sleep type (and what they would look like), eventually landing on the bear, the dolphin, the hedgehog, the giraffe, the parrot and the tortoise. “We feel connected to animals, and there aren’t often inherent negative associations with a dolphin or a hedgehog,” Elena explains. The Fitbit crew also didn’t want users to associate anything negative with their Sleep Profiles - they just wanted to better explain them. But in testing this and similar options, the team found that people had a hard time connecting to these representations. “We thought about using natural elements - like maybe a restless sleeper was a stormy ocean,” she says. That led to a bunch of brainstorming about different ways to represent the sleep types, says Elena Perez, a Fitbit product manager. “We needed to find a way to translate this sort of ‘mathy’ information about sleep types into something that’s more identifiable,” Karla says. To make this finding useful to people, Fitbit researchers began working with colleagues on Fitbit’s product side.
Then the team used machine learning technology to cluster the user data and found the answer to their first question: Yes, there are different types of sleepers - about six types, in fact.
Those sleep features included things like the probability of waking up in the first hour of sleep, sleep cycle length, wake time consistency and weekend versus weekday bedtime discrepancy. So the Fitbit team set out to accumulate some extremely useful data.įirst, they engineered more than 1,000 sleep features to capture users' sleeping behaviors in order to determine if there were sleeper types and what those types were. “We started with the research question of, ‘Are there different types of sleepers?’ There’s been a lot of interest in this question in academia, but not much capability to answer it until wearables.” But the team didn’t only want to figure out different sleep styles they also wanted to help people sleep better based on that information, Karla explains. “This was a multi-year effort,” says Fitbit research scientist Karla Gleichauf. Fitbit launched Sleep Profile last year as a Premium feature that gives users a detailed monthly analysis of their sleep after they’ve used the feature for at least 14 nights over a month.
0 kommentar(er)
