Location-targeted measurement of perceived stress

Abstract

Modern medicine has made significant advancements in the treatment of acute diseases, disorders, and injury. However, the acute care system rising from the twentieth century medical industry is far less adept in tracking and managing the needs of chronic conditions, especially that of chronic stress. Here we propose utilizing daily behavioral patterns to develop a model of daily stress, predictive of contexts leading to elevated perceived stress within an individual.

We identify the smartphone as an effective data logger of daily life, and demonstrate a battery-efficient method of all-day behavior tracking, particularly location and activity, of individuals within an open environment. This behavioral logging is then combined with on-device experience sampling to create and evaluate a model of perceived stress built from extracted features of daily activity, including exercise, social, and sleep.