High-Dimensional Visualization

High-dimensional data has many values for each data point and occurs frequently, including scientific data where lots of attributes are measured; engineering data capturing multiple sensor readings; and business data tracking different process metrics. Whether data comes from spreadsheets, databases or APIs, visualizing it collectively allows you to see overall patterns and trends between the various components.

Visualizing -dimensional data requires different techniques than for data with just 2 or 3 dimensions. Multipanel methods show multiple, simple views of the data using shared aesthetics and scales to make it easier to identify potential features. Single-panel plots use distinct markers and styles to represent the full depth of the data.

Pairwise Plots

Plot all pairwise 2D projections of data.

PairwiseListPlot array of pairwise list plots

PairwiseDensityHistogram array of pairwise density histograms

PairwiseSmoothDensityHistogram array of pairwise smooth density histograms

Compare all 1D projections of data.

PairwiseQuantilePlot array of pairwise Q-Q plots

PairwiseProbabilityPlot array of pairwise P-P plots

Multiaxis Plots

Plot an -dimensional point as a curve with multiple axes.

ParallelAxisPlot plot -dimensional points on parallel axes

RadialAxisPlot plot -dimensional points on radial axes

Feature Plots

Feature space plots extract low-dimensional feature vectors for each object and plot them.

FeatureSpacePlot plot feature vectors in 2D

FeatureSpacePlot3D plot feature vectors in 3D

Glyph Plots

Glyph plots visualize each data point as another plot.

PointValuePlot plot locations with associated data of any dimension