Interactive Analysis of Object Group Changes over Time

Abstract

The analysis of time-dependent data is an important task in various application domains. Often, the analyzed data objects belong to groups. The group memberships may stem from natural arrangements (e.g., animal herds), or may be constructed during analysis (e.g., by clustering). Group membership may change over time. Therefore, one important analytical aspect is to examine these changes (e.g., which herds change members and when). In this paper, we present a technique for visual analysis of group changes over time. We combine their visualization and automatic analysis. Interactive functions allow for tracking the data changes over time on group, set and individual level. We also consider added and removed objects (e.g., newly born or died animals). For large time series, automatic data analysis selects interesting time points and group changes for detailed examination. We apply our approach on the VAST 2008 challenge data set revealing new insights.