# Unobtrusive inference of diurnal rhythms from smartphone data

## TL;DR

This paper introduces a method to study daily rhythms using smartphone data, helping understand how sleep and behavior patterns change over time.

## Contribution

A novel digital phenotyping framework for quantifying diurnal rhythms using smartphone data is introduced.

## Key findings

- Smartphone typing dynamics can predict sleep duration effectively.
- Diurnal rhythm phase changes are detectable during time zone transitions.
- Longitudinal data supports the framework's utility in tracking behavioral patterns.

## Abstract

Diurnal rhythms are an integral feature of psychopathology but difficult to measure at scale. Smartphones are ubiquitous and therefore uniquely positioned to measure such rhythms non-invasively and continuously. Here, we propose a digital phenotyping framework to quantify diurnal rhythms. We use it to predict sleep duration from smartphone typing dynamics and analyse rhythm phase during time zone transitions with a clinical outpatient sample and a year-long longitudinal data set.