Search code examples
rplotmath

R: How do I put two box plots next to each other, keeping same y-range for both?


Let's say I have two data sets, one with y-range [min0:max0] and the other with y-range [min1:max1]. How can put both box plots in one plot next to each other with a sane y-range [min(min0, min1):max(max0, max1)]?

Here's what I tried:

d0 <- matrix(rnorm(15), ncol=3)
d1 <- matrix(rnorm(15), ncol=3)

par(mfrow = c(1, 2))
boxplot(d0)
usr <- par("usr")
plot.new()
par(usr = usr)
boxplot(d1, add = TRUE)

But that will only keep the first plots y-range and also squeeze both plots whereas I'd like them to be square.

Any ideas?


Solution

  • d0 <- matrix(rnorm(15), ncol=3)
    d1 <- matrix(rnorm(15), ncol=3)
    
    # Using base R graphics
    lmts <- range(d0,d1)
    
    par(mfrow = c(1, 2))
    boxplot(d0,ylim=lmts)
    boxplot(d1,ylim=lmts)
    

    base graphics plot

    You may also want to think about a way to do this using grid graphics, either by the lattice or ggplot2 packages.

    Here's one attempt in ggplot2:

    # Using ggplot2
    library(ggplot2)
    d <- data.frame(d.type=c(rep(0,15),rep(1,15)),sub.type=rep(c('A','B','C'),10),val=rnorm(30))
    
    p <- ggplot(d, aes(factor(sub.type), val)) 
    p + geom_boxplot() + facet_grid(. ~ d.type)
    

    ggplot2

    And in lattice:

    # Using lattice
    library(lattice)
    bwplot(~ val|sub.type+d.type ,d)
    

    lattice plot

    Note how the grid-based solutions keep you from ever having to specify limits; you specify structure and the software does the rest.