boxwork: Make boxplots
In metrumresearchgroup/pmplots: Plots for Pharmacometrics

boxwork

R Documentation

Make boxplots

Description

Make boxplots

Usage

boxwork(
  df,
  x,
  y,
  xs = defcx(),
  ys = defy(),
  fill = opts$boxplot.fill,
  alpha = opts$boxplot.alpha,
  hline = NULL,
  title = NULL,
  shown = TRUE,
  points = NULL,
  outlier.shape = opts$boxplot.outlier.shape,
  ...
)

Arguments

`df`	data frame to plot
`x`	character name for x-axis data
`y`	character name for y-axis data
`xs`	see `defcx`
`ys`	see `defy`
`fill`	passed to `geom_boxplot`
`alpha`	passed to `geom_boxplot`
`hline`	used to draw horizontal reference line
`title`	passed to `ggtitle`
`shown`	if `TRUE` provide a numeric summary of each box (see details)
`points`	show points in back of transparent boxes; if `TRUE`, a default display of points is made on top of boxes; also may be passed as a list of arguments to pass to `geom_point`; see details
`outlier.shape`	passed to `geom_boxplot`
`...`	arguments passed to `geom_boxplot`

Details

Since this function creates a boxplot, the x column must be character, factor or logical and y column must be numeric.

If shown is TRUE, a numeric summary of each box is included below each box. In the summary, n is the number of non-NA observations in the y column for that box and N is the number of unique ID values for that box. An error will be generated if ID does not exist in the plotting data frame when shown is TRUE. When N is equal to n in the summary, only n is shown.

The summaries will not be correct if the plot is eventually faceted by another variable in the data set. In this case, either use shown=FALSE or create the plot with split_plot.

When the user passes the points argument, outlier.shape is automatically switched to NA so that outlier points are only plotted once. The fill argument is also set to NA, so that boxes become transparent, showing the points.

When the user sets points to TRUE, grey points are shown in back of transparent boxes and the points are jittered in the x-direction. The user can customize the display of the points by passing a list of arguments for geom_point (for example, change the color, transparency, size, jitter amount, etc). The user can also pass jitter_width in the points list to set the amount of jitter in the x-direction while keeping jitter in the y-direction zero. Passing jitter_width will override any other value passed under position in the points list.