Visualization of data is valuable in statistical analysis to reveal features that are difficult to detect in numerical summaries. When p variables are measured on n subjects it can be difficult to graphically depict patterns due to limits of dimensionality. Multivariate graphical analysis identified directions that can be projected into subspaces to highlight sample features. Projecting the original data into a subspace generated from several orthogonal linear combinations allows for low-dimensional approximation to the full data representation.