Visualization Methods forExploratory Subgroup Discovery onTime Series Data

Abstract

This paper presents visualization methods for exploratory subgroup discovery, focusing on numeric time series data. We provide four novel visualizations for the inspection and understanding of subgroups. These visualizations facilitate interpretation in order to get insights into the data and the respective subgroups, while also supporting statistical interpretation and assessment of the subgroups and their respective parameters. Furthermore, we illustrate the approach in the context of complex time series data – specifically on team interactions in the affective computing context. © 2022, Springer Nature Switzerland AG.