CODEWITHSUNDEEP
rggplot2dendrogramhclust
Elizabeth
Elizabeth

Reputation: 6561

Adding labels to a dendrogram in ggplot using ggdendro in r

How can I add labels to a ggdendro plot? I realise from reading ?dendro_data that I am suppose to use the the call 'labels' but can't find an example of actual implementation. Could anybody please demonstrate how to add leaf labels to the example below. Thank you

require(ggplot2)

hc <- hclust(dist(USArrests), "ave")
dhc <- as.dendrogram(hc,hang=0.1)
ddata <- dendro_data(dhc, type="rectangle")
ggplot(segment(ddata)) + geom_segment(aes(x=x, y=y, xend=xend, yend=yend))

Upvotes: 1

Views: 5935

Answers (2)

Daniel Cook
Daniel Cook

Reputation: 43

I was able to plot hanging dendrograms in ggdendro without a substantial amount of work by applying the following. You just have to construct a dataframe of the labels that also contains the leaf positions. This is done by filtering out the points to only take integers.

# tree is a an extracted dendro_data() object.
label_data <- bind_cols(filter(segment(tree), x == xend & x%%1 == 0), label(tree))

ggplot() + 
geom_segment(data=segment(tree), aes(x=x, y=y, xend=xend, yend=yend)) +
geom_text(data=label_data, aes(x=xend, y=yend, label=label, hjust=0, color = LT), size=2) +
coord_flip() + 
scale_y_reverse(expand=c(0.2, 0)) +
theme_bw() +
theme(panel.border = element_blank(),
      panel.grid.major = element_blank(), 
      panel.grid.minor = element_blank(),
      axis.line = element_blank(),
      axis.title = element_blank(),
      axis.text = element_blank(),
      axis.ticks = element_blank(),
      legend.position = "None")

Upvotes: 0

orizon
orizon

Reputation: 3239

You can add leaf labels with a call to geom_text() using the data frame generated by label(ddata). I also extended the plot range using scale_y_continuous so the labels were not cutoff .

p <- ggplot(segment(ddata)) + geom_segment(aes(x=x, y=y, xend=xend, yend=yend))
p + geom_text(aes(x = x, y = y, label = label, angle = -90, hjust = 0), data= label(ddata)) +
  scale_y_continuous(expand = c(0.3, 0))

It may, however, be preferable to use ggdendrogram() unless you do not like the way those labels are displayed:

ggdendrogram(ddata)

Upvotes: 1

Related Questions