I would like to plot means and standard errors as a horizontal barchart, and I want the mean sorted.
I\'ve found the way to plot horizontal sorted barcharts using l
If it doesn't have to be lattice here is a simple function that uses base R functionality, that is supplied with three arguments: the widths of the bars (xv), the lengths (up and down) of the error bars (z) and the labels for the bars on the y axis (nn).
error.bars<-function(xv,z,nn){
par(las = 1)
yv <- barplot(xv,horiz = TRUE,col="cyan",xlim=c(0,(max(xv)+max(z))),names=nn,xlab=deparse(substitute(xv)))
g <- (max(yv)-min(yv))/(3*length(yv))
for (i in 1:length(yv)) {
lines(c(xv[i]+z[i],xv[i]-z[i]),c(yv[i],yv[i]))
lines(c(xv[i]+z[i],xv[i]+z[i]),c(yv[i]+g,yv[i]-g))
lines(c(xv[i]-z[i],xv[i]-z[i]),c(yv[i]+g,yv[i]-g))
}}
plot.data <- plot.data[order(plot.data$mean),] # reorder data
mean<-as.vector(plot.data$mean)
se<-as.vector(plot.data$error)
labels<-as.character(plot.data$var)
error.bars(mean,se,labels)