R adding a datatable to a ggplot graph using viewPorts : Scaling the Grob

Question

I'm trying to add a data-table to a graph made in ggplot (similar to the excel functionality but with the flexibility to change the axis its on)

I've had a few goes at it and keep hitting a problem with scaling so attempt 1) was

library(grid)
library(gridExtra)
library(ggplot2)
xta=data.frame(f=rnorm(37,mean=400,sd=50))
xta$n=0
for(i in 1:37){xta$n[i]<-paste(sample(letters,4),collapse='')}
xta$c=0
for(i in 1:37){xta$c[i]<-sample((1:6),1)}
rect=data.frame(xmi=seq(0.5,36.5,1),xma=seq(1.5,37.5,1),ymi=0,yma=10)
xta=cbind(xta,rect)
a = ggplot(data=xta,aes(x=n,y=f,fill=c)) + geom_bar(stat='identity')
b = ggplot(data=xta,aes(x=n,y=5,label=round(f,1))) + geom_text(size=4) + geom_rect(aes(xmin=xmi,xmax=xma,ymin=ymi,ymax=yma),alpha=0,color='black')
z = theme(axis.text=element_blank(),panel.background=element_rect(fill='white'),axis.ticks=element_blank(),axis.title=element_blank())
b=b+z
la=grid.layout(nrow=2,ncol=1,heights=c(0.15,2),default.units=c('null','null'))
grid.show.layout(la)
grid.newpage()
pushViewport(viewport(layout=la))
print(a,vp=viewport(layout.pos.row=2,layout.pos.col=1))
print(b,vp=viewport(layout.pos.row=1,layout.pos.col=1))

which produced

the second attempt 2) was

xta1=data.frame(t(round(xta$f,1)))
xtb=tableGrob(xta1,show.rownames=F,show.colnames=F,show.vlines=T,gpar.corefill=gpar(fill='white',col='black'),gp=gpar(fontsize=12),vp=viewport(layout.pos.row=1,layout.pos.col=1))
grid.newpage()
la=grid.layout(nrow=2,ncol=1,heights=c(0.15,2),default.units=c('null','null'))
grid.show.layout(la)
grid.newpage()
pushViewport(viewport(layout=la))
print(a,vp=viewport(layout.pos.row=2,layout.pos.col=1))
grid.draw(xtb)

which produced

and finally 3) was

grid.newpage()
print(a + annotation_custom(grob=xtb,xmin=0,xmax=37,ymin=450,ymax=460))

which produced

Of them option 2 would be the best if I could scale the tableGrob to the same size as the plot, but I've no idea how to do that. Any pointers on how to take this further? - Thanks

Answer 1

You can try the new version of tableGrob; the resulting gtable width/height can be set to a specific size (here equi-distributed npc units)

library(ggplot2)
library(gridExtra)
library(grid)
tg <- tableGrob(head(iris), rows=NULL)
tg$widths <- unit(rep(1/ncol(tg),ncol(tg)),"npc")
tg$heights <- unit(rep(1/nrow(tg),nrow(tg)),"npc")

qplot(colnames(iris), geom="bar")+ theme_bw() +
  scale_x_discrete(expand=c(0,0)) +
  scale_y_continuous(lim=c(0,2), expand=c(0,0)) +
  annotation_custom(ymin=1, ymax=2, xmin=-Inf, xmax=Inf, tg)

Answer 2

You can use for instance a table created by ggplot and combine them with like in this blog. I made a simplified and working example here:

First produce your plot:

library(ggplot2)
library(reshape2)
library(grid)

 df <- structure(list(City = structure(c(2L,
     3L, 1L), .Label = c("Minneapolis", "Phoenix",
     "Raleigh"), class = "factor"), January = c(52.1,
     40.5, 12.2), February = c(55.1, 42.2, 16.5),
     March = c(59.7, 49.2, 28.3), April = c(67.7,
         59.5, 45.1), May = c(76.3, 67.4, 57.1),
     June = c(84.6, 74.4, 66.9), July = c(91.2,
         77.5, 71.9), August = c(89.1, 76.5,
         70.2), September = c(83.8, 70.6, 60),
     October = c(72.2, 60.2, 50), November = c(59.8,
         50, 32.4), December = c(52.5, 41.2,
         18.6)), .Names = c("City", "January",
     "February", "March", "April", "May", "June",
     "July", "August", "September", "October",
     "November", "December"), class = "data.frame",
     row.names = c(NA, -3L))

dfm <- melt(df, variable = "month")

 levels(dfm$month) <- month.abb
 p <- ggplot(dfm, aes(month, value, group = City,
     colour = City))
 p1 <- p + geom_line(size = 1) + theme(legend.position = "top") + xlab("")

Next produce the data table in ggplot. Use the same x-axis as the plot:

none <- element_blank()
data_table <- ggplot(dfm, aes(x = month, y = factor(City),
     label = format(value, nsmall = 1), colour = City)) +
     geom_text(size = 3.5) +
  scale_y_discrete(labels = abbreviate)+ theme_bw()  +
     theme(panel.grid.major = none, legend.position = "none",
         panel.border = none, axis.text.x = none,
         axis.ticks = none) + theme(plot.margin = unit(c(-0.5,
     1, 0, 0.5), "lines")) + xlab(NULL) + ylab(NULL)

Combine the two with viewport:

Layout <- grid.layout(nrow = 2, ncol = 1, heights = unit(c(2,
     0.25), c("null", "null")))
grid.show.layout(Layout)
vplayout <- function(...) {
     grid.newpage()
     pushViewport(viewport(layout = Layout))
 }

subplot <- function(x, y) viewport(layout.pos.row = x,
     layout.pos.col = y)

mmplot <- function(a, b) {
     vplayout()
     print(a, vp = subplot(1, 1))
     print(b, vp = subplot(2, 1))
 }

mmplot(p1, data_table)

Note that still some tweaking is needed like the position of the legend of the plot and the abbrevation of the city names in the table, but the result looks nice:

Applied to your example:

library(grid)
library(gridExtra)
library(ggplot2)
xta=data.frame(f=rnorm(37,mean=400,sd=50))
xta$n=0
for(i in 1:37){xta$n[i]<-paste(sample(letters,4),collapse='')}
xta$c=0
for(i in 1:37){xta$c[i]<-sample((1:6),1)}
rect=data.frame(xmi=seq(0.5,36.5,1),xma=seq(1.5,37.5,1),ymi=0,yma=10)
xta=cbind(xta,rect)
a = ggplot(data=xta,aes(x=n,y=f,fill=c)) + geom_bar(stat='identity')+ theme(legend.position = "top")+xlab("")

none <- element_blank()
z=ggplot(xta, aes(x = n, y = "fvalues",
     label = round(f,1)) )+
     geom_text(size = 3)+ theme_bw()  +
     theme(panel.grid.major = none, legend.position = "none",
         panel.border = none, axis.text.x = none,
         axis.ticks = none) + theme(plot.margin = unit(c(-0.5,
     1, 0, 0.5), "lines")) + xlab(NULL) + ylab(NULL)

Layout <- grid.layout(nrow = 2, ncol = 1, heights = unit(c(2,
     0.25), c("null", "null")))
grid.show.layout(Layout)
vplayout <- function(...) {
     grid.newpage()
     pushViewport(viewport(layout = Layout))
 }

subplot <- function(x, y) viewport(layout.pos.row = x,
     layout.pos.col = y)

mmplot <- function(a, b) {
     vplayout()
     print(a, vp = subplot(1, 1))
     print(b, vp = subplot(2, 1))
 }

mmplot(a, z)

EDIT:

similar to Dennis his solution but than a barplot and with + coord_flip(). You can remove the latter if you don't want to flip it, but it increases readability:

ggplot(xta, aes(x=n,y=f,fill=c)) +
   geom_bar() +
   labs(color = "c") +
   geom_text(aes(y = max(f)+30, label = round(f, 1)), size = 3, color = "black") + coord_flip()

Answer 3

IMHO this is not a well designed graphic. Firstly, I don't understand why you need a zero origin when the values range from 300 to 500, which is a disguised way of saying that I don't like the bar chart metaphor. You're also attempting to use bar fill to represent differences in the values of c, of which there are only six. Here's what I believe is a simpler approach to the problem. Given your xta data,

# Convert the categories to a factor
xta$N <- factor(xta$n, levels = xta$n)

# Simple approach:
ggplot(xta, aes(x = f, y = N, color = factor(c))) +
   geom_point() +
   labs(color = "c") +
   geom_text(aes(x = 575, label = round(f, 1)), size = 4, color = "black")

That's not an interesting graphic to me. What might add a little insight, depending on the context of the problem, would be to sort the responses in increasing order and add a size aesthetic to punctuate differences between levels of c. (You could also use size without color.) Finally, because we put the factor levels on the vertical axis so that their labels are clearly visible, we can also insert the f values as text by extending the horizontal axis a bit.

ggplot(xta, aes(x = f, y = reorder(N, f), color = factor(c), size = c)) + 
   geom_point() +
   labs(color = "c") +
   geom_text(aes(x = 575, label = round(f, 1)), size = 4, color = "black")

There are enough hints in this code for you to take it in a different direction. I'll leave that up to you.