Integration of statistical computing and interactive visualization in a single framework

Abstract

The challenge of analyzing heterogeneous data sets to find patterns, draw conlusions, or simply get a better understanding of how the data is structured is a central topic of statistics. Visualization is a vital tool to understand data, its effectiveness relying on the great perceptual abilities of the human brain.<br />Recent developments show that data sets tend to be increasingly complex and come from various sources, ranging from business applications over scientific research studies to simulation results in engineering. Accordingly, different types of variables (continuous, categorical, ordinal, etc.) are encountered which, combined with high dimensionality and large sample sizes, continuously increase the computational power needed for satisfactory data visualization and analysis.<br />Still, data analysis cannot solely rely on visual exploration of datasets, and more sophisticated statistical procedures are necessary.<br />To allow for a comprehensive view on data, it is therefore pertinent to integrate both visual and computational methods in a single framework.