ABSTRACT

The Scagnostics system provides a scalable summary of large scatterplot matrices through a derived scatterplot matrix (SPLOM) based on classifying the geometric shapes formed by the point distributions in the original scatterplots. The Vis DB system for database vis treats an entire database as a very large table of data, visually encoding it with dense and space-filling layouts with items colored according to their relevance for a specific query. The Hierarchical Clustering Explorer system supports systematic exploration of multidimensional tables, such as those representing microarray measurements of gene expression in the genomics domain, with an associated hierarchical clustering. The ability to concisely describe existing systems gives a firm foundation for considering the full array of possibilities when generate new systems. Graph-theoretic scagnostics is a scalable idiom for the exploration of scatterplot matrices, or SPLOMs. A scagnostics SPLOM is a next step beyond a standard SPLOM, just as a SPLOM is a step beyond a single scatterplot.