Multivariate Data Display

Several techniques have been proposed to display multivariate data.

• Axis reconfiguration techniques, such as parallel coordinates [11,22] and glyphs [3,6,19].
• Dimensional embedding techniques, such as dimensional stacking [17], trellis displays [5] and worlds within worlds [8].
• Dimensional subsetting, such as scatterplots [7].
• Dimensional reduction techniques, such as multidimensional scaling [1,16,24], principal component analysis [13] and self-organizing maps [15].

Problem

Most of these techniques do not scale well with respect to size.

Our Objectives

• Interactive visualization of large multivariate datasets on parallel coordinates.
• Using a hierarchical approach that presents a multiresolutional view of the data.
• Suite of navigation tools to facilitate the systematic discovery of data trends and hidden patterns.