For sleep researchers, this paper is interesting for two reasons. One, the authors have contributed a useful new tool called SleepExplorer that correlates multiple contextual measurements and even introduces a new structure for organizing sleep measurements. Two, perhaps even more important than the tool, they’ve taken a bold stance regarding sleep research to date: the metrics aren’t usable. Now, they don’t state it as baldly as that; they maintain a proper academic tone and demeanor throughout the paper. However, it’s quite clear that most sleep quality indices to date are “magic numbers” that vary from tool to tool, researcher by researcher, and simply aren’t comprehensive enough to be very useful. The new reporting structure and visualization tool introduced by the authors are intended to extend and standardize sleep measurement, leading to improved understanding and reporting.
But sleep researchers aren’t the only ones who might find this paper useful. Computing in general is undergoing a fundamental change in becoming the Internet of Things (IoT). To date, sensors (things) have been single purposed, transmitting scalar values (temperature, humidity, photos, and so on) to the cloud for analysis. There’s a growing awareness, though, that single values are far less useful than environmental/situational context (known as “sensor fusion” in IoT). That’s an exact parallel to the conclusion of this paper: context is king. It’s rewarding to recognize concepts that cross disciplines.