A LARGE SCALE ANALYSIS OF PHONE USAGE PATTERNS TO UNDERSTAND RHYTHMS OF HUMAN ACTIVITIES
Open Access
- Author:
- Jaeger, Ryan Edward
- Area of Honors:
- Information Sciences and Technology
- Degree:
- Bachelor of Science
- Document Type:
- Thesis
- Thesis Supervisors:
- Saeed Abdullah, Thesis Supervisor
David Joseph Fusco, Thesis Honors Advisor - Keywords:
- Circadian Computing
Smartphones
Social jet lag
Data science
Device Analyzer
Sleep timing
Mobogram
Health informatics - Abstract:
- Smartphones are becoming increasingly prevalent around the world, and time spent on these devices is rising each year. For that reason, smartphones offer particularly useful insights into the timing of human activity, and have become an alternative approach to tracking sleep behaviors in individuals. Social jet lag - the difference in sleep timing between work days and free days - is a prevalent sleep disruption and has been associated with a number of health risks. Previous studies of social jet lag been small in scale or based on self-reported sleep information, which can be unreliable. Here we show how smartphone usage patterns can be harnessed to quantify social jet lag in a global population. To do so, we used a dataset on mobile device usage events called Device Analyzer collected by researcher at the University of Cambridge. The dataset has over 27,000 devices from countries around the world, and therefore this study is the first smartphone-based study of social jet lag with global implications. We found that smartphone usage patterns can be useful sources of information on human activity as they reflect daily routines. Additionally, we found that the median social jet lag across a global population of Android users was approximately 60 minutes, and over 50% of individuals experience social jet lags greater than 1 hour, indicating social jet lag to be a widespread disruption. Our quantification of social jet lag aligns with previous studies, and suggests that the phone use patterns could be a potentially rich source of information for future circadian computing research.