Reprojecting visualizations for advanced interaction

Abstract

Visual analytics (VA) is increasingly important in data exploration and analysis. While the qualitative results of interactive visual analysis (IVA) remain an essential strength of VA, we believe that extending the current IVA approach is necessary to support critical applications such as medical diagnosis and decision-making. This master thesis supports the existing research by incorporating more quantitative results and allowing users to reproduce their brushing results. Overlaid descriptive statistics and the relative difference plot contribute to decision-making and interpretation. The structured brushing space and novel brushes like percentile and Mahalanobis enable interactive and reproducible quantitative analysis. Positive feedback validates its applicability across domains.