TimeSeriesPlot {FitAR} | R Documentation |
Cleveland pointed out that the aspect-ratio is important in graphically showing the rate-of-change or shape information. For many time series, it is preferably to set this ratio to 0.25 than the default. In general, Cleveland shows that the best choice of aspect-ratio is often obtained by if the average apparent absolute slope in the graph is about 45 deg. But for many stationary time series, this would result in an aspect-ratio which would be too small. As a comprise we have chosen a default of 0.25 but the user can select other choices.
TimeSeriesPlot(z, SubLength = Inf, aspect = 0.25, type="l", ylab=NULL, main=NULL, ...)
z |
ts object or vector, time series data |
SubLength |
maximum number of data points per panel. Default SubLength=Inf and regular graphics. For trellis graphics, set SubLength to a finite value. |
aspect |
optional setting for the aspect-ratio |
type |
plot type, default type="l" join points with lines |
ylab |
optional label for vertical axis |
main |
optional title |
... |
optional arguments passed to xyplot |
If z has attribute "title" containing a character string, this is used
on the plot.
Time series input using the function Readts
always
have this attribute set.
If SubLength
is finite, the lattice package is used and a graphic object of class trellis is produced.
Otherwise, the standard R graphics system is used and the plot is produced as a side-effect and there is no output.
Requires lattice
library
A.I. McLeod
W.S. Cleveland (1993), Visualizing Data.
#from built-in datasets TimeSeriesPlot(AirPassengers) title(main="Monthly number of trans-Atlantic airline passengers") # #compare plots for lynx series plot(lynx) win.graph() TimeSeriesPlot(lynx, type="o", pch=16, ylab="# pelts", main="Lynx Trappings") # #lattice style plot data(Ninemile) TimeSeriesPlot(Ninemile, SubLength=200)