B_03_histogram {lattice}R Documentation

Histograms and Kernel Density Plots


Draw Histograms and Kernel Density Plots, possibly conditioned on other variables.


histogram(x, data, ...)
densityplot(x, data, ...)
## S3 method for class 'formula'
          allow.multiple, outer = TRUE,
          auto.key = FALSE,
          aspect = "fill",
          panel = lattice.getOption("panel.histogram"),
          prepanel, scales, strip, groups,
          xlab, xlim, ylab, ylim,
          type = c("percent", "count", "density"),
          nint = if (is.factor(x)) nlevels(x)
          else round(log2(length(x)) + 1),
          endpoints = extend.limits(range(as.numeric(x),
                          finite = TRUE), prop = 0.04),
          equal.widths = TRUE,
          drop.unused.levels =
          lattice.options = NULL,
          default.scales = list(),
          default.prepanel =

## S3 method for class 'numeric'
histogram(x, data = NULL, xlab, ...)
## S3 method for class 'factor'
histogram(x, data = NULL, xlab, ...)

## S3 method for class 'formula'
            allow.multiple = is.null(groups) || outer,
            outer = !is.null(groups),
            auto.key = FALSE,
            aspect = "fill",
            panel = lattice.getOption("panel.densityplot"),
            prepanel, scales, strip, groups, weights,
            xlab, xlim, ylab, ylim,
            bw, adjust, kernel, window, width, give.Rkern,
            n = 512, from, to, cut, na.rm,
            drop.unused.levels =
            lattice.options = NULL,
            default.scales = list(),
            default.prepanel =
## S3 method for class 'numeric'
densityplot(x, data = NULL, xlab, ...)

do.breaks(endpoints, nint)



The object on which method dispatch is carried out.

For the formula method, x can be a formula of the form ~ x | g1 * g2 * ..., indicating that histograms or kernel density estimates of the x variable should be produced conditioned on the levels of the (optional) variables g1, g2, .... x should be numeric (or possibly a factor in the case of histogram), and each of g1, g2, ... should be either factors or shingles.

As a special case, the right hand side of the formula can contain more than one term separated by ‘+’ signs (e.g., ~ x1 + x2 | g1 * g2). What happens in this case is described in the documentation for xyplot. Note that in either form, all the terms in the formula must have the same length after evaluation.

For the numeric and factor methods, x is the variable whose histogram or Kernel density estimate is drawn. Conditioning is not allowed in these cases.


For the formula method, an optional data source (usually a data frame) in which variables are to be evaluated (see xyplot for details). data should not be specified for the other methods, and is ignored with a warning if it is.


A character string indicating the type of histogram that is to be drawn. "percent" and "count" give relative frequency and frequency histograms respectively, and can be misleading when breakpoints are not equally spaced. "density" produces a density histogram.

type defaults to "density" when the breakpoints are unequally spaced, and when breaks is NULL or a function, and to "percent" otherwise.


An integer specifying the number of histogram bins, applicable only when breaks is unspecified or NULL in the call. Ignored when the variable being plotted is a factor.


A numeric vector of length 2 indicating the range of x-values that is to be covered by the histogram. This applies only when breaks is unspecified and the variable being plotted is not a factor. In do.breaks, this specifies the interval that is to be divided up.


Usually a numeric vector of length (number of bins + 1) defining the breakpoints of the bins. Note that when breakpoints are not equally spaced, the only value of type that makes sense is density.

When breaks is unspecified, the value of lattice.getOption("histogram.breaks") is first checked. If this value is NULL, then the default is to use

      breaks = seq_len(1 + nlevels(x)) - 0.5

when x is a factor, and

      breaks = do.breaks(endpoints, nint)

otherwise. Breakpoints calculated in such a manner are used in all panels. If the retrieved value is not NULL, or if breaks is explicitly specified, it affects the display in each panel independently. Valid values are those accepted as the breaks argument in hist. In particular, this allows specification of breaks as an integer giving the number of bins (similar to nint), as a character string denoting a method, or as a function.

When specified explicitly, a special value of breaks is NULL, in which case the number of bins is determined by nint and then breakpoints are chosen according to the value of equal.widths.


A logical flag, relevant only when breaks=NULL. If TRUE, equally spaced bins will be selected, otherwise, approximately equal area bins will be selected (typically producing unequally spaced breakpoints).


Integer, giving the number of points at which the kernel density is to be evaluated. Passed on as an argument to density.


A function, called once for each panel, that uses the packet (subset of panel variables) corresponding to the panel to create a display. The default panel functions panel.histogram and panel.densityplot are documented separately, and have arguments that can be used to customize its output in various ways. Such arguments can usually be directly supplied to the high-level function.

allow.multiple, outer

See xyplot.


See xyplot.


See xyplot.


See xyplot.


See xyplot.


See xyplot.


See xyplot. Note that the default panel function for histogram does not support grouped displays, whereas the one for densityplot does.

xlab, ylab

See xyplot.

xlim, ylim

See xyplot.


See xyplot.


See xyplot.


See xyplot.


See xyplot.


See xyplot.


Fallback prepanel function. See xyplot.


numeric vector of weights for the density calculations, evaluated in the non-standard manner used for groups and terms in the formula, if any. If this is specified, it is subsetted using subscripts inside the panel function to match it to the corresponding x values.

At the time of writing, weights do not work in conjunction with an extended formula specification (this is not too hard to fix, so just bug the maintainer if you need this feature).

bw, adjust, width

Arguments controlling bandwidth. Passed on as arguments to density.

kernel, window

The choice of kernel. Passed on as arguments to density.


Logical flag, passed on as argument to density. This argument is made available only for ease of implementation, and will produce an error if TRUE.

from, to, cut

Controls range over which density is evaluated. Passed on as arguments to density.


Logical flag specifying whether NA values should be ignored. Passed on as argument to density, but unlike in density, the default is TRUE.


Further arguments. See corresponding entry in xyplot for non-trivial details.


histogram draws Conditional Histograms, and densityplot draws Conditional Kernel Density Plots. The default panel function uses the density function to compute the density estimate, and all arguments accepted by density can be specified in the call to densityplot to control the output. See documentation of density for details.

These and all other high level Trellis functions have several arguments in common. These are extensively documented only in the help page for xyplot, which should be consulted to learn more detailed usage.

do.breaks is an utility function that calculates breakpoints given an interval and the number of pieces to break it into.


An object of class "trellis". The update method can be used to update components of the object and the print method (usually called by default) will plot it on an appropriate plotting device.


The form of the arguments accepted by the default panel function panel.histogram is different from that in S-PLUS. Whereas S-PLUS calculates the heights inside histogram and passes only the breakpoints and the heights to the panel function, lattice simply passes along the original variable x along with the breakpoints. This approach is more flexible; see the example below with an estimated density superimposed over the histogram.


Deepayan Sarkar Deepayan.Sarkar@R-project.org


Sarkar, Deepayan (2008) Lattice: Multivariate Data Visualization with R, Springer. http://lmdvr.r-forge.r-project.org/

See Also

xyplot, panel.histogram, density, panel.densityplot, panel.mathdensity, Lattice


histogram( ~ height | voice.part, data = singer, nint = 17,
          endpoints = c(59.5, 76.5), layout = c(2,4), aspect = 1,
          xlab = "Height (inches)")

histogram( ~ height | voice.part, data = singer,
          xlab = "Height (inches)", type = "density",
          panel = function(x, ...) {
              panel.histogram(x, ...)
              panel.mathdensity(dmath = dnorm, col = "black",
                                args = list(mean=mean(x),sd=sd(x)))
          } )

densityplot( ~ height | voice.part, data = singer, layout = c(2, 4),  
            xlab = "Height (inches)", bw = 5)

[Package lattice version 0.20-38 Index]