TypeAware : collecting smartphone typing and sleep data to promote digital well-being

Abstract

Is it possible to harvest all the data our smartphones and wearable sleep trackers provide to build recommendations that improve our digital well-being and sleeping habits? This thesis contributes to the ongoing research on sleep-recommending platforms by developing a mobile application named TypeAware. Kotlin programming language is used to build the application and implement features where the user's typing speed, error rate and current time of day can be recorded and visualised to be presented to the user in an informative manner. Additionally, Fitbit's API is used to obtain the sleep data displayed in the application.

TypeAware is used to conduct a user study where participants are recruited from Prolific to use the application for 11 days per participant. Every day the participants are presented with a questionnaire to study how they perceive the smartphone usage and sleep data provided by the application and whether they can gain meaningful insights from their data during the study period.

The study revealed that people find it easier to gain insights from sleep-related rather than smartphone usage data. Sleep metrics are also considered more useful when making recommendations to others on how to improve their sleep or digital well-being. In the end half of the study participants thought that the data provided by the TypeAware application would encourage them to make behavioural changes and could also prove useful for other people. It could not be proved that the amount of sleep would affect people's ability to better estimate their smartphone usage. There were some indications that people who sleep more are also typing on their phones less.