Systematic Categorisation of 3,091 Smartphone Applications From a Large-Scale Smartphone Sensing Dataset

Abstract

Practically all user activities on a smartphone depend on self-contained software applications, so-called apps. Due to the large number and diversity of available apps, the analysis of app usage behaviour in social science research requires elaborate pre-processing of app data. Therefore, we present a categorisation scheme and a dataset of 3,091 manually categorised apps used by a representative quota sample within a large-scale smartphone sensing study conducted in Germany over several months in 2020. For the categorisation, we report values for inter-rater agreement between two independent raters. We provide the freely available dataset as a CSV and we invite other researchers to use and modify the categorisation for their specific research questions and to extend it for the mobile sensing research community.