CODEWITHSUNDEEP
rggplot2
dpwr
dpwr

Reputation: 2812

Add axis labels to tableGrob in R

When using the default table output, the default is to plot the "axis" labels using the parameter names (k1 and k2 in this case). This is useful particularly for comparing cluster membership and the most important thing to denote is which cluster is rows and which is columns.

table(
  k1=matrix(1:4, 2),
  k2=matrix(1:4, 2)
)

   k2
k1  1 2 3 4
  1 1 0 0 0
  2 0 1 0 0
  3 0 0 1 0
  4 0 0 0 1

I am trying to make use of tableGrob to output some tables, composed with other ggplot components in grid.arrange. However, the axis labels are lost.

grid.arrange(tableGrob(
  table(
    k1=matrix(1:4, 2),
    k2=matrix(1:4, 2)
  )
))

missingAxis

All I want to do is have those included or even include them manually after making the grob.

Thanks

Edit: Attempt with annotation_custom:

annotation_custom

Code:

ggTableAxis2 <- function(t) {
  my_grob <- tableGrob(t)

  my_plot <- ggplot(mapping = aes(k2, k1)) + 
    annotation_custom(my_grob) +
    scale_x_continuous(position = 'top') +
    theme(axis.title =  element_text(angle = 0, hjust = 0),
          axis.title.x.top =  element_text(hjust = 0))
  return(my_plot)
}

grid.arrange(
  ggTableAxis2(
    table(
      k1=matrix(1:4, 2),
      k2=matrix(1:4, 2)
    )
  ),
  ggTableAxis2(
    table(
      k1=matrix(1:4, 2),
      k2=matrix(1:4, 2)
    )
  ),
  nrow=1
)

Upvotes: 2

Views: 1805

Answers (2)

baptiste
baptiste

Reputation: 77096

not sure where exactly you want the labels, but it's probably easiest to add grobs to the gtable,

tbl <- table(
  longlabel1=matrix(1:4, 2),
  longlabel2=matrix(1:4, 2)
)

nms <- names(dimnames(tbl))


library(gridExtra)
library(grid)
library(gtable)
grid.newpage()

tg <- textGrob(nms[1], x=0, hjust=0)
lg <- textGrob(nms[2], x=1, hjust=1)
g <- tableGrob(tbl)
g <- gtable_add_rows(g, unit(1,"line"), 0)
g <- gtable_add_cols(g, grobWidth(lg), 0)
g <- gtable_add_grob(g, tg, t = 1, l=3, r=ncol(g))
g <- gtable_add_grob(g, lg, l = 1, t=2)
grid.draw(g)

enter image description here

Upvotes: 5

Axeman
Axeman

Reputation: 35187

You could use an empty ggplot with mapped aesthetics:

library(grid)
library(gridExtra)
library(ggplot2)

my_grob <- tableGrob(table(k1 = matrix(1:4, 2), k2 = matrix(1:4, 2)))

my_plot <- ggplot(mapping = aes(k2, k1)) + 
  annotation_custom(my_grob) +
  scale_x_continuous(position = 'top') +
  theme(axis.title.y =  element_text(angle = 0),
        axis.title.x.top =  element_text(hjust = 0))

grid.arrange(my_plot, my_plot, my_plot, my_plot)

enter image description here

Upvotes: 2

Related Questions