When users' devices are online or offline during the day, they follow the daily cycle of human behavior: a trough in the early hours, followed by a spike in activity during the day to a peak in the evening, then a sharp drop overnight.
However, no two cycles are the same: day of the week is important (heading downtown on a Friday night lowers internet activity), stay-at-home requests certainly matter (we get online earlier and longer), and even a dip in activity during prayer times during a month Ramadan visible in traditional Islamic areas.
The American Time Use Survey (ATUS) asks Americans about their previous day's activities, including when they woke up and when they went to sleep. The Monash study used survey data of 81 US cities over a six-year period to calculate when residents sleep and wake up each year, and then used Internet activity data to do the same calculation.
The researchers then trained a machine-learning algorithm to track how changes in internet use over the course of a day correlate with average wake and sleep times in each city. minute. When estimating the average morning wake-up time, it was accurate to within nine minutes.
The researchers repeated this finding when using daily electricity demand data instead of internet data to predict sleep. But there is something fundamentally different about measures of internet activity compared to electricity demand data: global availability.
The United States has a high-performance electrical bureaucracy, but not all countries do. On the other hand, telemetry and continual measurement of internet activity can be done for any internet-connected device on the planet, and this suggests that the amount of sleep we get collectively can be estimated for any A city (connected to the internet) on the planet in almost real time.
This type of research has a wide range of applications, including impact mapping during natural disasters, documenting internet shutdowns linked to human rights abuses, and even providing internet availability assessments during the Russo-Ukrainian war.