splot

Splot is a term found in several disciplines, most notably in computer plotting and in statistical data analysis. In computing, splot commonly refers to a command in the plotting program gnuplot for 3D surface plotting, while in statistics and particle physics it denotes the sPlot technique for statistically separating signal and background components.

In gnuplot, the splot command generates three-dimensional plots, typically surfaces, as opposed to the 2D plot command. It can plot surfaces from data files or from mathematical functions, and supports styling options such as lines, points, or filled surfaces, as well as various projection and coloring schemes. A typical usage is splot 'data.dat' using 1:2:3 with lines, which renders a surface defined by columns 1, 2, and 3.

Statistical sPlot (often capitalized SPlot) refers to a weighting method introduced by Pivk and Le Diberder to extract signal distributions from data containing multiple components. It uses a maximum-likelihood fit to a discriminating variable (for example, an invariant mass) to obtain component yields and computes per-event weights, sWeights, to reconstruct distributions of other variables for each component. The method assumes the discriminating variable is uncorrelated with the variables of interest for each component; violations of this assumption can bias results.

Other uses are rare and vary by domain; in some contexts splot may be used as a generic shorthand for surface plotting in three-dimensional visualization. See also: gnuplot, sPlot.