triplot: A single plot with 4 variables
In ptarroso/assortedViz: Assorted functions for plotting

Description Usage Arguments Details Value Author(s) References See Also Examples

View source: R/triplot.R

Plots 4 variables in a single plot. Two independent variables are plotted horizontally and a third one is plotted vertically, allowing to display data in a scatter plot. Data from the fourth variables is extracted at the points within user-defined squares in the plotted area and aggregated with a function (default average). Aggregated data is displayed with different color accordingly to the color scale defined. Squares are divided in two triangles with the top one referring to values aggregated in relation to the variable on top axis, and lower triangle in relation to bottom axis variable.

triplot(x1, x2, y, z, size = 0.025, ps = 1, FUN = mean, col.FUN = hcl.colors, breaks = 100, x1.at = NULL, x1.lab = x1.at, x2.at = NULL, x2.lab = x2.at, y.at = NULL, y.lab = y.at, displayNA = TRUE, axis.labels = NULL, scale = TRUE, scale.at=NULL, scale.lab=scale.at, verbose=TRUE, ...)

`x1`	A vector of data for the bottom variable.
`x2`	A vector of data (same length as x1) for the top variable.
`y`	A vector of data (same length as x1) for the vertical variable.
`z`	A vector of data (same length as x1) to be aggregated for the x1 vs. y and for x2 vs. y relations.
`size`	Size of square to aggregate data. See description.
`ps`	Size of plotted squares/triangles.
`FUN`	Function to aggregate data (default is mean)
`col.FUN`	Function for color gradient (default hcl.colors).
`breaks`	Number of color breaks of sequence of breaks.
`x1.at`	Locations for tick marks on bottom axis.
`x1.lab`	Labels for tick marks on bottom axis.
`x2.at`	Locations for tick marks on top axis.
`x2.lab`	Labels for tick marks on top axis.
`y.at`	Locations for tick marks on left axis.
`y.lab`	Labels for tick marks on left axis.
`displayNA`	Display squares with NA values?
`axis.labels`	A vector of length 4 with axis labels for X1, X2, Y and Z.
`scale`	Add a color scale bar for Z values? If TRUE, than a multiplot is created with a scale bar on bottom.
`scale.at`	Position of tick marks on scale bar (if null - default - five equally spaced labels are added).
`scale`	Labels for each tick mark in the scale bar.
`verbose`	If TRUE (default) it displays a progress bar during calculations.
`...`	Additional arguments for plotting.

This function allows to plot four variables. Originally was intended for plotting variables that shared a spatial location (for instance, spatial raster data) but can be used with other type of data.

Both X variables and Y are internally scaled to the range form zero to one for having a full squared plotting area. The plotting area is divided in smaller squares of side of length *size*. If the *size* argument is one, than only one square is fitted.

Two independent scatter plots are computed internally: one is X1 vs. Y and the other is X2 vs. Y. The values of Z are found for the points within each smaller square and aggregated with *FUN* (default to mean value). The aggregated value for the X1/Y is displayed in the bottom triangle of the divided square and the value for X2/Y in the top triangle. Values are color coded accordingly to the color function given.

A plot!

If the function is redirected to a variable 'dt <- triplot(...)' it returns a list with 3 items:

1. tri - the value of the triangle in respect to x1 vs. y and x2 vs. y.

2. original.crd - a 3 column matrix with real values for the plot grid for each of the varaibles.

3. plot.crd - a 2 column table with grid values on the plot coordinates (from zero to one). Since x1 = x2 in this case, only x is reported.

Pedro Tarroso

(recursive reference)

tile

	# Just demonstration of simple functionality
	triplot(iris$Sepal.Length, iris$Petal.Length, iris$Sepal.Width, iris$Petal.Width)

	# With a slightly better display
	lbls <- c("Sepal Length", "Petal Length", "Sepal Width", "Petal Width")
	triplot(iris$Sepal.Length, iris$Petal.Length, iris$Sepal.Width, iris$Petal.Width,
          size=0.1, displayNA=F, border="white",
          x1.at = c(5, 6, 7), x2.at=c(2, 4, 6), y.at=c(2, 3, 4),
          axis.labels=lbls)