Consider the following situation:
we have quantified algal concentrations for
a variety of species using many samples at each
of three years. It seems to make sense to generate
a line plot (matplot-like), with each species plotted
as a separate line, with the points connected to emphasize
the temporal pattern.
The problem: lots of overlapping error bars.
The question: from both a procedural and an aesthetic
point of view, what do people recommend? Below is an
example that hacks (not quite "jitters", because it's
systematic) the x locations a bit, but I find it a bit
ugly. (I was also hoping that ggplot had some magic
to do this automatically, but position_dodge doesn't
seem to work in this context ...)
Ben Bolker
library(ggplot2)
nx <- 3
ngrp <- 5
nper <- 4
x <- rep(1:nx,ngrp*nper)
y <- runif(nx*ngrp*nper)
g <- factor(rep(1:ngrp,each=nx*nper))
dat <- data.frame(x,y,g)
se <- with(dat,tapply(y,list(g,x),sd)/table(g,x))
means <- with(dat,tapply(y,list(g,x),mean))
limits <- aes(ymax=means+se,ymin=means-se)
gg <- factor(rep(1:ngrp,nx))
xx <- rep(1:nx,each=ngrp)
p <- ggplot(dat,aes(y=c(means),x=xx,
group=gg,colour=gg))
p + geom_pointrange(aes(ymin=c(means-se),ymax=c(means+se)))
xdodge <- xx + rep(seq(-0.1,0.1,length=ngrp),nx)
p + geom_line() + geom_pointrange(aes(x=xdodge,
ymin=c(means-se),
ymax=c(means+se))) +
scale_x_continuous(name="Year",breaks=1:3) +
scale_y_continuous(name="Whatever")
--
Ben Bolker
Associate professor, Biology Dep't, Univ. of Florida
bol...@ufl.edu / www.zoology.ufl.edu/bolker
GPG key: www.zoology.ufl.edu/bolker/benbolker-publickey.asc
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
