SPLOM

SPLOM (scatterplot matrix) is a graphical display used in exploratory data analysis to show pairwise relationships among multiple quantitative variables. In a SPLOM each row and column corresponds to a variable, and each off-diagonal cell contains a scatterplot of the variable for its row against the variable for its column. Diagonal cells often display a univariate summary such as a histogram, density estimate, boxplot, or variable name.

The technique helps analysts detect correlations, nonlinear relationships, clusters, and outliers without imposing a model. It is particularly useful for small- to medium-dimensional datasets where visual comparison of many two-dimensional projections can reveal structure that summary statistics might miss. Common enhancements include adding regression or smoothing lines, annotating cells with correlation coefficients, and using color or symbol shape to encode a categorical variable.

Limitations of SPLOMs include rapid growth in the number of plots as variable count increases (O(n^2)), difficulty reading detailed structure when datasets have many points due to overplotting, and sensitivity to scale differences among variables. Practical workarounds include variable selection, sampling, transparency or binning techniques (e.g., hexbin), and interactive linked brushing. SPLOMs are widely implemented in statistical and visualization software (for example R, Python libraries, and commercial tools) and appear under related names such as pair plot, pairwise scatterplot matrix, and corner plot.