Falling for ggplot2

I spent a long time resisting the lure of ggplot2. I was proficient with the plotting functions in base graphics; why did I need to learn an entirely new graphics system? Yes, getting up colour ramps could be a real pain, but I was in complete control.

Recently I’ve had to learn how to use ggplot2. Students were asking me to help them fix their code (often copied off Stack Overflow), so I needed at least some understanding. I read the book – eller hente ebok frå Universitetsbiblioteket. Its very readable. Now I see why ggplot2 is so popular and have started to use it whenever possible.

Today I wanted to make a screeplot for a principal component analysis in ggplot. So I wrote my own function based on screeplot in the vegan package.

ggscreeplot <-
function (x, bstick = FALSE, npcs = min(10, if (is.null(x$CCA) || x$CCA$rank == 0) x$CA$rank else x$CCA$rank),
          xlab = "Component", ylab = "Inertia", title = "",  ...) {
  if (is.null(x$CCA) || x$CCA$rank == 0)
    eig.vals <- x$CA$eig
  else eig.vals <- x$CCA$eig
  ncomps <- length(eig.vals)   if (npcs > ncomps)
    npcs <- ncomps
  comps <- seq(len = npcs)

  df <- data.frame(Inertia = eig.vals[comps], names = reorder(factor(names(eig.vals[comps])), comps))
  ymax <- df$Inertia[1]   if (bstick && !is.null(x$CCA) && x$CCA$rank > 0) {
    warning("'bstick' unavailable for constrained ordination")
    bstick <- FALSE
  } else {
    df <- cbind(df, bstick = bstick(x)[comps])
    ymax <-max(ymax, df$bstick[1])
  g <- ggplot(df, aes(x = names)) +
                geom_bar(aes(y = Inertia), stat = "identity", fill = "grey50") +
                labs(x = xlab, y = ylab, title = title) +
                scale_y_continuous(limits = c(0, ymax * 1.04), expand = c(0, 0))

  if (bstick) {
                g <- g + geom_point(aes(y = bstick), colour = "red")
                g <- g + geom_line(aes(y = bstick, group =1), colour = "red")

Here it is in action

data(RLGH)#Round Loch of Glenhead diatom stratigraphy
ggscreeplot(rda(sqrt(RLGH$spec)), bstick = TRUE)

Screeplot of a PCA of the Round Loch of Glenhead diatom stratigraphy. Note that axis one is much longer than the other axes – it has almost half the total variance.

I’ll be using this code when I show ordinations of the Lake Żabińskie chironomid reconstruction. Spoiler – they don’t look much like this.


About richard telford

Ecologist with interests in quantitative methods and palaeoenvironments
This entry was posted in EDA, R. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s