Produce a Sunflower Scatter Plot

sunflowerplot {graphics}

R Documentation

Produce a Sunflower Scatter Plot

Description

Multiple points are plotted as ‘sunflowers’ with multiple leaves (‘petals’) such that overplotting is visualized instead of accidental and invisible.

Usage

sunflowerplot(x, ...)

## Default S3 method:
sunflowerplot(x, y = NULL, number, log = "", digits = 6,
              xlab = NULL, ylab = NULL, xlim = NULL, ylim = NULL,
              add = FALSE, rotate = FALSE,
              pch = 16, cex = 0.8, cex.fact = 1.5,
              col = par("col"), bg = NA, size = 1/8, seg.col = 2,
              seg.lwd = 1.5, ...)

## S3 method for class 'formula'
sunflowerplot(formula, data = NULL, xlab = NULL, ylab = NULL, ...,
             subset, na.action = NULL)

Arguments

`x`	numeric vector of `x`-coordinates of length `n`, say, or another valid plotting structure, as for `plot.default`, see also `xy.coords`.
`y`	numeric vector of `y`-coordinates of length `n`.
`number`	integer vector of length `n`. `number[i]` = number of replicates for `(x[i], y[i])`, may be 0. Default (`missing(number)`): compute the exact multiplicity of the points `x[], y[]`, via `xyTable()`.
`log`	character indicating log coordinate scale, see `plot.default`.
`digits`	when `number` is computed (i.e., not specified), `x` and `y` are rounded to `digits` significant digits before multiplicities are computed.
`xlab, ylab`	character label for x-, or y-axis, respectively.
`xlim, ylim`	`numeric(2)` limiting the extents of the x-, or y-axis.
`add`	logical; should the plot be added on a previous one ? Default is `FALSE`.
`rotate`	logical; if `TRUE`, randomly rotate the sunflowers (preventing artefacts).
`pch`	plotting character to be used for points (`number[i]==1`) and center of sunflowers.
`cex`	numeric; character size expansion of center points (s. `pch`).
`cex.fact`	numeric shrinking factor to be used for the center points when there are flower leaves, i.e., `cex / cex.fact` is used for these.
`col, bg`	colors for the plot symbols, passed to `plot.default`.
`size`	of sunflower leaves in inches, 1[in] := 2.54[cm]. Default: 1/8\", approximately 3.2mm.
`seg.col`	color to be used for the segments which make the sunflowers leaves, see `par(col=)`; `col = "gold"` reminds of real sunflowers.
`seg.lwd`	numeric; the line width for the leaves' segments.
`...`	further arguments to `plot` [if `add = FALSE`], or to be passed to or from another method.
`formula`	a `formula`, such as `y ~ x`.
`data`	a data.frame (or list) from which the variables in `formula` should be taken.
`subset`	an optional vector specifying a subset of observations to be used in the fitting process.
`na.action`	a function which indicates what should happen when the data contain `NA`s. The default is to ignore case with missing values.

Details

This is a generic function with default and formula methods.

For number[i] == 1, a (slightly enlarged) usual plotting symbol (pch) is drawn. For number[i] > 1, a small plotting symbol is drawn and number[i] equi-angular ‘rays’ emanate from it.

If rotate = TRUE and number[i] >= 2, a random direction is chosen (instead of the y-axis) for the first ray. The goal is to jitter the orientations of the sunflowers in order to prevent artefactual visual impressions.

Value

A list with three components of same length,

`x`	x coordinates
`y`	y coordinates
`number`	number

Use xyTable() (from package grDevices) if you are only interested in this return value.

Side Effects

A scatter plot is drawn with ‘sunflowers’ as symbols.

Author(s)

Andreas Ruckstuhl, Werner Stahel, Martin Maechler, Tim Hesterberg, 1989–1993. Port to R by Martin Maechler maechler@stat.math.ethz.ch.

References

Chambers, J. M., Cleveland, W. S., Kleiner, B. and Tukey, P. A. (1983). Graphical Methods for Data Analysis. Wadsworth.

Schilling, M. F. and Watkins, A. E. (1994). A suggestion for sunflower plots. The American Statistician, 48, 303–305. doi: 10.2307/2684839.

Murrell, P. (2005). R Graphics. Chapman & Hall/CRC Press.

Examples

require(stats) # for rnorm
require(grDevices)

## 'number' is computed automatically:
sunflowerplot(iris[, 3:4])
## Imitating Chambers et al, p.109, closely:
sunflowerplot(iris[, 3:4], cex = .2, cex.fact = 1, size = .035, seg.lwd = .8)
## or
sunflowerplot(Petal.Width ~ Petal.Length, data = iris,
              cex = .2, cex.fact = 1, size = .035, seg.lwd = .8)


sunflowerplot(x = sort(2*round(rnorm(100))), y = round(rnorm(100), 0),
              main = "Sunflower Plot of Rounded N(0,1)")
## Similarly using a "xyTable" argument:
xyT <- xyTable(x = sort(2*round(rnorm(100))), y = round(rnorm(100), 0),
               digits = 3)
utils::str(xyT, vec.len = 20)
sunflowerplot(xyT, main = "2nd Sunflower Plot of Rounded N(0,1)")

## A 'marked point process' {explicit 'number' argument}:
sunflowerplot(rnorm(100), rnorm(100), number = rpois(n = 100, lambda = 2),
              main = "Sunflower plot (marked point process)",
              rotate = TRUE, col = "blue4")

[Package graphics version 3.5.0 Index]