Sum Up: Statistical Visual Query of Multivariate Data with Parallel-Coordinate Geometry

摘要

One of the most noticeable issues of parallel coordinate visualization is how to quantitatively analyze density caused by polyline growth in a limited space on axes. The existing visualization tools only support the comparison among single dimensions and single ranges of polylines, which could face limitation in cases of complicated analytics. This paper proposes a new visual-query technique, named SumUp, for statistical analysis of multiple attributes of dimensions and multiple ranges of polylines. The methodology of SumUp is primarily based on developing dynamic queries using brushing operations to deliver summary stacked bars adaptive with parallel coordinates. Users can easily observe quantitative information from data patterns and compare multiple attributes over the density of polylines in the parallel coordinate visualization. Early experiments show that our proposed technique could potentially enhance the manipulation on parallel coordinates, showing by a typical case study.