Add contours to a ternary plot

Draws contour lines to depict the value of a function in ternary space.

TernaryContour(
  Func,
  resolution = 96L,
  direction = getOption("ternDirection", 1L),
  region = getOption("ternRegion", ternRegionDefault),
  within = NULL,
  filled = FALSE,
  legend,
  legend... = list(),
  nlevels = 10,
  levels = pretty(zlim, nlevels),
  zlim,
  color.palette = function(n) hcl.colors(n, palette = "viridis", alpha = 0.6),
  fill.col = color.palette(length(levels) - 1),
  func... = list(),
  ...
)

Arguments

Func: Function taking vectors of coordinates a, b and c, which returns a numeric vector whose value at each coordinate will be depicted.
resolution: The number of triangles whose base should lie on the longest axis of the triangle. Higher numbers will result in smaller subdivisions and smoother colour gradients, but at a computational cost.
direction: (optional) Integer specifying the direction that the current ternary plot should point: 1, up; 2, right; 3, down; 4, left.
region: (optional) Named list of length two specifying the the minimum and maximum values of each ternary axis to be drawn (e.g. list(min = c(40, 0, 0), max = c(100, 60, 60)); or a set of coordinates in a format accepted by TernaryPoints(). The plotted region will correspond to the smallest equilateral triangle that encompasses the specified ranges or coordinates.
within: List or matrix of x, y coordinates within which contours should be evaluated, in any format supported by xy.coords(x = within). If NULL, defaults to a region slightly smaller than the ternary plot. The $hull entry generated by TriangleInHull() may also be used.
filled: Logical; if TRUE, contours will be filled (using .filled.contour().).
legend: Character vector specifying annotations for colour scale. If not provided, no colour legend is displayed. Specify TRUE to generate automatically, or a single integer to generate legend annotations.
legend...: List of additional parameters to send to SpectrumLegend().
nlevels, levels, zlim, ...: parameters to pass to contour().
color.palette: parameters to pass to .filled.contour().
fill.col: Sent as col parameter to .filled.contour(). Computed from color.palette if not specified.
func...: List of additional parameters to send to Func().

Value

TernaryContour() invisibly returns a list containing:

x,y: the Cartesian coordinates of each evaluated point;
z: The value of Func() at each coordinate.

Author

Martin R. Smith (martin.smith@durham.ac.uk)

Examples

FunctionToContour <- function (a, b, c) {
  a - c + (4 * a * b) + (27 * a * b * c)
}

# Set up plot
originalPar <- par(mar = rep(0, 4))
TernaryPlot(alab = "a", blab = "b", clab = "c")
values <- TernaryPointValues(FunctionToContour, resolution = 24L)
ColourTernary(
  values,
  legend = signif(seq(max(values), min(values), length.out = 4), 2),
  bty = "n"
)
TernaryContour(FunctionToContour, resolution = 36L)


# Note that FunctionToContour is sent a vector.
# Instead of
BadMax <- function (a, b, c) {
  max(a, b, c)
}

# Use
GoodMax <- function (a, b, c) {
  pmax(a, b, c)
}
TernaryPlot(alab = "a", blab = "b", clab = "c")
ColourTernary(TernaryPointValues(GoodMax))
TernaryContour(GoodMax)


# Or, for a generalizable example,
GeneralMax <- function (a, b, c) {
  apply(rbind(a, b, c), 2, max)
}
TernaryPlot(alab = "a", blab = "b", clab = "c")
# Fill the contour areas, rather than using tiles
TernaryContour(GeneralMax, filled = TRUE,
               legend = c("Max", "...", "Min"), legend... = list(bty = "n"),
               fill.col =  hcl.colors(14, palette = "viridis", alpha = 0.6))
# Re-draw edges of plot triangle over fill
TernaryPolygon(diag(3))


# Restore plotting parameters
par(originalPar)

Arguments

Value

See also

Author

Examples