QS18 Preview: Esther Dyson and Three Sleep Trackers
September 5, 2018
Esther Dyson is a board member of 23andme, former chair of ICANN and the Electronic Frontier Foundation, and an investor in companies like Omada Health, PatientsLikeMe, and Medspace. But, like the rest of us, she spends a good portion of her life unconscious. While sleeping, she collects data with three different devices: Oura, Whoop, and ResMed. Each device tracks slightly differently: the Oura tracks sleep via a ring worn on the finger, while the Whoop is a wrist-based tracker and the ResMed S+ sits near the bed and uses radio waves to detect body movement. Unsurprisingly, the devices don’t always agree.
In her upcoming Show&Tell talk at QS18, Esther will talk about why she uses three sleep tracking devices and how she interprets their disparate outputs, providing a close look at the type of questions that interest many of us when we’re starting a self-tracking project. What constitutes “accurate” data? What are the differences between supposedly reliable tools? From the ill-fated Zeo to the S+, Whoop, and Oura, sleep has been a notoriously difficult metric to track. Whether it is because of an inaccurate sensor or the difficulty of distinguishing sleep patterns of the target individual from surrounding noise, sleep has been much tougher to track than many other metrics. Dyson’s talk provides insight into how to deal with those difficulties and still create useful insights from sleep data.
By looking at data from multiple sensors, Dyson gives a picture of how data can change according to the instrument used to record it, but also points to a central feature of a successful self-tracking project: that is, to create useful, actionable conclusions in the context of one person’s life, answering questions not about the universally verifiable accuracy of sleep measurement, but about her own sleep.
You can check out the full schedule of show&tell talks, breakout discussions and how-to workshops at the 2018 Quantified Self Conference in Portland, Oregon.