Star Plots of Multivariate Data

DESCRIPTION:

Creates star plots from a matrix of multivariate data. Options are available to control the type of star produced and the layout of the stars.

USAGE:

stars(x, full=T, scale=T, radius=T, type="l", labels=<<see below>>,  
      head=<<see below>>, max=<<see below>>, byrow=F,  
      nrow=<<see below>>, ncol=<<see below>>)

REQUIRED ARGUMENTS:

x: matrix of data. One star symbol will be produced for each row of the matrix. Missing values ( NAs) are allowed, but they are treated as if they were 0.

OPTIONAL ARGUMENTS:

full

logical flag: if TRUE, the symbols will occupy a full circle. Otherwise, they occupy the (upper) semi-circle only.

scale

logical flag: if TRUE, the columns of the data matrix are scaled independently so that the maximum value in each column is 1 and the minimum is 0. If FALSE, the presumption is that the data have been scaled by some other algorithm to the range `0<=x[i,j]<=1'.

radius

logical flag: if TRUE, the radii corresponding to each variable in the data will be drawn.

type

character string, giving the type of star to draw. Reasonable values are "l", "p", "b" for lines, points and both.

labels

vector of character strings for labeling the plots. If omitted, the first component of dimnames(x) is used, if available; otherwise, labels are "1", "2", etc. If supplied, the label vector must have length equal to nrow(x).

head

character string for a running head. This is plotted as a title at the top of each page. By default, the name of the data is used. If there is more than one page, the page number is included in the title.

max

suggested limit for the number of rows or columns of plots on a single page. The algorithm tries to choose an array of plots which efficiently uses the available space on the display. The default forces all symbols to be on one page.

byrow

logical flag: should the symbols be plotted row-by-row across the page, or column-by-column?

nrow,ncol

may be given to specify exactly the number of rows and columns for the array of plots on each page.

Graphical parameters may also be supplied as arguments to this function (see par ).

SIDE EFFECTS:

one or more pages of star plots are created on the current graphics device.

DETAILS:

Missing values are treated as 0.

Each star represents one row of the input x. Variables (columns) start on the right and wind counterclockwise around the star. The size of a (scaled) column is shown by the distance from the center to the point on the star representing the variable.

EXAMPLES:

# The stars() function is roughly equivalent to the following 
# function, which you can customize. 
my.stars <- 
function(x, labels = dimnames(x)[[1]], nrow = 
    floor(sqrt(dim(x)[1])), ncol = ceiling(dim(x)[1]/nrow)) 
{ 
    opar <- par(mar = c(2, 0, 0, 0), mfcol = c(nrow, ncol)) 
    on.exit(par(opar)) 
    for(i in 1:dim(x)[1]) { 
        starsymb(x, collab = rep("", dim(x)[2]), sample = i) 
        if(!is.null(labels)) 
            mtext(text = labels[i], side = 1, line = 0, cex = 1) 
    } 
} 
stars(votes.repub[state.region==1,]/100, radius=T, scale=F, 
      head="Republican Votes (Northeast) 1856 - 1976")