p.ts: plot.ts with multi-plots and Auto-Title -- on 1 page

Description

For longer time-series, it is sometimes important to spread the time-series plots over several subplots. p.ts(.) does this both automatically, and under manual control.

Actually, this is a generalization of plot.ts (with different defaults).

Usage

p.ts(x, nrplots = max(1, min(8, n %/% 400)), overlap = nk %/% 16,
     date.x = NULL, do.x.axis = !is.null(date.x), do.x.rug = FALSE,
     ax.format, main.tit = NULL, ylim = NULL, ylab = "", xlab = "Time",
     quiet = FALSE, mgp = c(1.25, .5, 0), ...)

Arguments

x: timeseries (possibly multivariate) or numeric vector.
nrplots: number of sub-plots. Default: in {1..8}, approximately n/400 if possible.
overlap: by how much should subsequent plots overlap. Defaults to about 1/16 of sub-length on each side.
date.x: a time “vector” of the same length as x and coercable to class "POSIXct" (see DateTimeClasses).
do.x.axis: logical specifying if an x axis should be drawn (i.e., tick marks and labels).
do.x.rug: logical specifying if rug of date.x values should drawn along the x axis.
ax.format: when do.x.axis is true, specify the format to be used in the call to axis.POSIXct.
main.tit: Main title (over all plots). Defaults to name of x.
ylim: numeric(2) or NULL; if the former, specifying the y-range for the plots. Defaults to a common pretty range.
ylab, xlab: labels for y- and x-axis respectively, see description in plot.default.
quiet: logical; if TRUE, there's no reporting on each subplot.
mgp: numeric(3) to be passed to mult.fig(), see par(mgp = .).
...: further graphic parameters for each plot.ts(..).

Side Effects

A page of nrplots subplots is drawn on the current graphics device.

Author

Martin Maechler, maechler@stat.math.ethz.ch; July 1994 (for S).

Examples

Run this code

stopifnot(require(stats))
## stopifnot(require(datasets))

data(sunspots)
p.ts(sunspots, nr=1) # == usual  plot.ts(..)
p.ts(sunspots)
p.ts(sunspots, nr=3, col=2)

data(EuStockMarkets)
p.ts(EuStockMarkets[,"SMI"])
## multivariate :
p.ts(log10(EuStockMarkets), col = 2:5)

## with Date - x-axis (dense random dates):
set.seed(12)
x <- as.Date("2000-02-29") + cumsum(1+ rpois(1000, lambda= 2.5))
z <- cumsum(.1 + 2*rt(1000, df=3))
p.ts(z, 4, date.x = x)
p.ts(z, 6, date.x = x, ax.format = "%b %Y", do.x.rug = TRUE)