Combining grid_arrange_shared_legend() and facet_wrap_labeller() in R - r

I am trying to combine grid_arrange_shared_legend() and facet_wrap_labeller() in R. More specifically, I want to draw a figure including two ggplot figures with multiple panels each and have a common legend. I further want to italicize part of the facet strip labels. The former is possible with the grid_arrange_shared_legend() function introduced here, and the latter can be achieved with the facet_wrap_labeller() function here. However, I have not been successful in combining the two.
Here's an example.
library("ggplot2")
set.seed(1)
d <- data.frame(
f1 = rep(LETTERS[1:3], each = 100),
f2 = rep(letters[1:3], 100),
v1 = runif(3 * 100),
v2 = rnorm(3 * 100)
)
p1 <- ggplot(d, aes(v1, v2, color = f2)) + geom_point() + facet_wrap(~f1)
p2 <- ggplot(d, aes(v1, v2, color = f2)) + geom_smooth() + facet_wrap(~f1)
I can place p1 and p2 in the same figure and have a common legend using grid_arrange_shared_legend() (slightly modified from the original).
grid_arrange_shared_legend <- function(...) {
plots <- list(...)
g <- ggplotGrob(plots[[1]] + theme(legend.position = "right"))$grobs
legend <- g[[which(sapply(g, function(x) x$name) == "guide-box")]]
lheight <- sum(legend$width)
grid.arrange(
do.call(arrangeGrob, lapply(plots, function(x)
x + theme(legend.position = "none"))),
legend,
ncol = 2,
widths = unit.c(unit(1, "npc") - lheight, lheight))
}
grid_arrange_shared_legend(p1, p2)
Here's what I get.
It is possible to italicize part of the strip label by facet_wrap_labeller().
facet_wrap_labeller <- function(gg.plot,labels=NULL) {
require(gridExtra)
g <- ggplotGrob(gg.plot)
gg <- g$grobs
strips <- grep("strip_t", names(gg))
for(ii in seq_along(labels)) {
modgrob <- getGrob(gg[[strips[ii]]], "strip.text",
grep=TRUE, global=TRUE)
gg[[strips[ii]]]$children[[modgrob$name]] <- editGrob(modgrob,label=labels[ii])
}
g$grobs <- gg
class(g) = c("arrange", "ggplot",class(g))
g
}
facet_wrap_labeller(p1,
labels = c(
expression(paste("A ", italic(italic))),
expression(paste("B ", italic(italic))),
expression(paste("C ", italic(italic)))
)
)
However, I cannot combine the two in a straightforward manner.
p3 <- facet_wrap_labeller(p1,
labels = c(
expression(paste("A ", italic(italic))),
expression(paste("B ", italic(italic))),
expression(paste("C ", italic(italic)))
)
)
p4 <- facet_wrap_labeller(p2,
labels = c(
expression(paste("A ", italic(italic))),
expression(paste("B ", italic(italic))),
expression(paste("C ", italic(italic)))
)
)
grid_arrange_shared_legend(p3, p4)
# Error in plot_clone(p) : attempt to apply non-function
Does anyone know how to modify either or both of the functions so that they can be combined? Or is there any other way to achieve the goal?

You need to pass the gtable instead of the ggplot,
library(gtable)
library("ggplot2")
library(grid)
set.seed(1)
d <- data.frame(
f1 = rep(LETTERS[1:3], each = 100),
f2 = rep(letters[1:3], 100),
v1 = runif(3 * 100),
v2 = rnorm(3 * 100)
)
p1 <- ggplot(d, aes(v1, v2, color = f2)) + geom_point() + facet_wrap(~f1)
p2 <- ggplot(d, aes(v1, v2, color = f2)) + geom_smooth() + facet_wrap(~f1)
facet_wrap_labeller <- function(g, labels=NULL) {
gg <- g$grobs
strips <- grep("strip_t", names(gg))
for(ii in seq_along(labels)) {
oldgrob <- getGrob(gg[[strips[ii]]], "strip.text",
grep=TRUE, global=TRUE)
newgrob <- editGrob(oldgrob,label=labels[ii])
gg[[strips[ii]]]$children[[oldgrob$name]] <- newgrob
}
g$grobs <- gg
g
}
combined_fun <- function(p1, p2, labs1) {
g1 <- ggplotGrob(p1 + theme(legend.position = "right"))
g2 <- ggplotGrob(p2 + theme(legend.position = "none"))
g1 <- facet_wrap_labeller(g1, labs1)
legend <- gtable_filter(g1, "guide-box", trim = TRUE)
g1p <- g1[,-(ncol(g1)-1)]
lw <- sum(legend$width)
g12 <- rbind(g1p, g2, size="first")
g12$widths <- unit.pmax(g1p$widths, g2$widths)
g12 <- gtable_add_cols(g12, widths = lw)
g12 <- gtable_add_grob(g12, legend,
t = 1, l = ncol(g12), b = nrow(g12))
g12
}
test <- combined_fun(p1, p2, labs1 = c(
expression(paste("A ", italic(italic))),
expression(paste("B ", italic(italic))),
expression(paste("C ", italic(italic)))
)
)
grid.draw(test)

Related

R shared legend error

I'm using a shared legend in R, which is using this code:
g <- ggplotGrob(plots[[1]] + theme(legend.position="bottom"))$grobs
legend <- g[[which(sapply(g, function(x) x$name) == "guide-box")]]
Reference: Combining grid_arrange_shared_legend() and facet_wrap_labeller() in R
What is "structure(c(3L,3L,……))" in my legend? How to avoid it?
My code:
library(RColorBrewer)
library(ggplot2)
library(gridExtra)
wq <- read.csv('wineQualityReds.csv')
quality_factor <- factor(wq$quality)
plot_three <- function(feature) {
ggplot(data=wq,
aes_string(x = feature, y = wq$pH, color = quality_factor)) +
geom_tile() +
scale_color_brewer(palette='Blues') +
geom_smooth(method = loess, aes(group = 1)) + theme(legend.position = "none")
}
ph1 <- plot_three("volatile.acidity") +
labs(x='volatile.acidity(g / dm^3)', y='pH')
ph2 <- plot_three("citric.acid") +
labs(x='citric.acid(g / dm^3)', y='pH')
ph3 <- plot_three("chlorides") +
labs(x='chlorides(g / dm^3)', y='pH')
ph4 <- plot_three("free.sulfur.dioxide") +
labs(x='free.sulfur.dioxide(mg / dm^3)', y='pH')
ph5 <- plot_three("sulphates") +
labs(x='sulphates(g / dm^3)', y='pH')
ph6 <- plot_three("alcohol") +
labs(x='alcohol(% by volume)', y='pH')
g <- ggplotGrob(ph1 + theme(legend.position="bottom"))$grobs
legend <- g[[which(sapply(g, function(x) x$name) == "guide-box")]]
grid.arrange(arrangeGrob(ph1, ph2, ph3, ph4, ph5, ph6, ncol =3),
legend, top = "Relationship between pH and features via different quality", heights=c(9,1))
Dataset: redwineQuality
https://s3.amazonaws.com/udacity-hosted-downloads/ud651/wineQualityReds.csv

Align gridArranged facetted ggplots

I've created a faceted plot, separately for three different groups in my data, like so:
df <- data.frame(x=rep(seq(0.05,1,by=0.05),times=40),
y=sample(c('A','B'),20*40,replace=TRUE),
id=rep(1:40,each=20),
group=c(rep(1,20*12),rep(2,20*12),rep(3,20*16)))
g1 <- ggplot(df[df$group==1,],aes(x,y,group=id))
g1 <- g1 + geom_line()
g1 <- g1 + facet_wrap(~id,ncol=3)
g2 <- ggplot(df[df$group==2,],aes(x,y,group=id))
g2 <- g2 + geom_line()
g2 <- g2 + facet_wrap(~id,ncol=3)
g3 <- ggplot(df[df$group==3,],aes(x,y,group=id))
g3 <- g3 + geom_line()
g3 <- g3 + facet_wrap(~id,ncol=3)
grid.arrange(g1,g2,g3,nrow=1)
which gives me this:
As you can see, the number of facets differs between the three groups which means that the facets in the three columns have different heights. Is there any way to harmonize this height in a non-fragile way (i.e. without me having to manually determine the heights of columns 2 and 3 that gives me facets that look like they have roughly the same height)?
Here's a solution with some guidance from this question.
library(ggplot2)
library(gridExtra)
ncol = 3
df <- data.frame(x=rep(seq(0.05,1,by=0.05),times=40),
y=factor(sample(c('A','B'),20*40,replace=TRUE), levels = c("A", "B")),
id=rep(1:40,each=20),
group=c(rep(1,20*12),rep(2,20*12),rep(3,20*16)))
max_cases <- max(table(unique(df[,c("id", "group")])$group))
# create phantom plots for everything in the containing rectangle to standardize labels
rect_dim <- ceiling(max_cases / ncol) * ncol
plots <- lapply(X=unique(df$group), FUN= function(i){
df_case <- subset(df, subset= group == i)
tot_case <- nrow(unique(df_case[,c("id", "group")]))
# create fill levels to pad the plots
fill_levels <- unlist(lapply(X=1:(rect_dim - tot_case), function(y){paste0(rep(x=" ", times=y), collapse="")}))
df_case$id.label <- ordered(df_case$id, levels = c(unique(df_case$id), fill_levels))
g_case <- ggplot(df_case,aes(x,y,group=id.label)) +
geom_line() +
facet_wrap(~id.label, ncol = ncol, drop=FALSE)
# whiteout the inner y axis elements to clean it up a bit
if(i != 1){
g_case <- g_case + theme(axis.text.y = element_text(color = "white"),
axis.title.y = element_text(color = "white"),
axis.ticks.y = element_line(color = "white"))
}
g_case <- ggplotGrob(g_case)
rm_me <- (tot_case:rect_dim)[-1]
# remove empty panels and layout
g_case$grobs[names(g_case$grobs) %in% c(paste0("panel", rm_me), paste0("strip_t.", rm_me))] <- NULL
g_case$layout <- g_case$layout[!(g_case$layout$name %in% c(paste0("panel-", rm_me), paste0("strip_t-", rm_me))),]
g_case
})
plots$nrow = 1
do.call("grid.arrange", plots)
It's a bit messy, but you can massage the gtables to have the same number of rows, and align them. Further refinement would locate the rows corresponding to plot panels, rather than assume that all plots have the same row sequence of panel - axes - etc.
library(gtable)
cbind_top = function(...){
pl <- list(...)
## test that only passing plots
stopifnot(do.call(all, lapply(pl, inherits, "gg")))
gl <- lapply(pl, ggplotGrob)
nrows <- sapply(gl, function(x) length(x$heights))
tallest <- max(nrows)
add_dummy <- function(x, n){
if(n == 0) return(x)
gtable_add_rows(x, rep(unit(0, "mm"), n), nrow(x)-2)
}
gl <- mapply(add_dummy, x=gl, n=tallest - nrows)
compare_unit <- function(u1,u2){
n <- length(u1)
stopifnot(length(u2) == n)
null1 <- sapply(u1, attr, "unit")
null2 <- sapply(u2, attr, "unit")
null12 <- null1 == "null" | null2 == "null"
both <- grid::unit.pmax(u1, u2)
both[null12] <- rep(list(unit(1,"null")), sum(null12))
both
}
bind2 <- function(x,y){
y$layout$l <- y$layout$l + ncol(x)
y$layout$r <- y$layout$r + ncol(x)
x$layout <- rbind(x$layout, y$layout)
x$widths <- gtable:::insert.unit(x$widths, y$widths)
x$colnames <- c(x$colnames, y$colnames)
x$heights <- compare_unit(x$heights, y$heights)
x$grobs <- append(x$grobs, y$grobs)
x
}
combined <- Reduce(bind2, gl[-1], gl[[1]])
grid::grid.newpage()
grid::grid.draw(combined)
}
cbind_top(g1,g2,g3)

ggplot2:scatterplots for all possible combinations of variables

I want to plot graphs for all possible combinations of variables. My code is below:
set.seed(12345)
a <- data.frame(Glabel=LETTERS[1:7], A=rnorm(7, mean = 0, sd = 1), B=rnorm(7, mean = 0, sd = 1), C=rnorm(7, mean = 0, sd = 1))
T <- data.frame(Tlabel=LETTERS[11:20], A=rnorm(10, mean = 0, sd = 1), B=rnorm(10, mean = 0, sd = 1), C=rnorm(10, mean = 0, sd = 1))
library(ggplot2)
for(i in 2:(ncol(a)-1))
{
for(j in (i+1):ncol(a))
{
r <- 0.08
p <- ggplot(data=a, mapping=aes(x=a[, i], y=a[, j])) + geom_point() + theme_bw()
p <- p + geom_text(data=a, mapping=aes(x=a[, i], y=a[, j], label=Glabel),
size=3, vjust=1.35, colour="black")
p <- p + geom_segment(data = T, aes(xend = T[ ,i], yend=T[ ,j]),
x=0, y=0, colour="black",
arrow=arrow(angle=25, length=unit(0.25, "cm")))
p <- p + geom_text(data=T, aes(x=T[ ,i], y=T[ ,j], label=Tlabel), size=3, vjust=0, colour="red")
dev.new()
print(p)
}
}
This code works fine. But the method used here is not recommended (See #baptiste comment) and does not work in function. I want to know what is the best and recommended way to accomplish this task. Thanks in advance for your help.
Alright this is garbage but the best I could do. It's super inefficient as it recreates a partial data with each loop through lapply. Maybe someone else has something better:
MAT <- outer(names(df)[-1], names(df)[-1], paste)
combs <- sapply(MAT[lower.tri(MAT)], function(x) strsplit(x, " "))
ind <- lapply(combs, function(x) match(x, names(df)))
plotter <- function(cn) { #start junky function
NAMES <- colnames(df)[cn]
df2 <- df[cn]
names(df2)<- c('x1', 'x2')
p <- ggplot(data=df2, aes(x1, x2)) + geom_point() + theme_bw() +
scale_x_continuous(name=NAMES[1]) +
scale_y_continuous(name=NAMES[2])
dev.new()
print(p)
} #end of junky function
lapply(ind, function(x) plotter(cn=x))
EDIT: This is a bit better:
x <- match(names(df)[-1], names(df))
MAT <- outer(x, x, paste)
combs <- t(sapply(MAT[lower.tri(MAT)], function(x) as.numeric(unlist(strsplit(x, " ")))))
plotter <- function(cn) {
NAMES <- colnames(df)[cn]
df2 <- df[cn]
names(df2)<- c('x1', 'x2')
p <- ggplot(data=df2, aes(x1, x2)) + geom_point() + theme_bw() +
scale_x_continuous(name=NAMES[1]) +
scale_y_continuous(name=NAMES[2])
dev.new()
print(p)
}
apply(combs, 1, function(x) plotter(cn=x))

ggplot2 boxplot: horizontal bar at median?

I would like to make a ggplot2 boxplot more meaningful by adding a thick bar at the median (so that if the median is equal to either of the lower or upper quartiles, it can be detected to which it is equal). I came across a recent post of Kohske:
Can I get boxplot notches in ggplot2?
but I didn't know how to give the "crossbar" a "height". Then I tried
to use a rectangle but it didn't work either. Here is a minimal example:
require(ggplot2)
require(reshape2)
require(plyr)
set.seed(1)
## parameters
p1 <- c(5, 20, 100)
p2 <- c("f1", "f2", "f3", "f4", "f5")
p3 <- c("g1","g2","g3","g4","g5")
N <- 1000
## lengths
l1 <- length(p1)
l2 <- length(p2)
l3 <- length(p3)
## build result array containing the measurements
arr <- array(rep(NA, l1*l2*l3*N), dim=c(l1, l2, l3, N),
dimnames=list(
p1=p1,
p2=p2,
p3=p3,
N=1:N))
for(i in 1:l1){
for(j in 1:l2){
for(k in 1:l3){
arr[i,j,k,] <- i+j+k+runif(N, min=-4, max=4)
}
}
}
arr <- arr + rexp(3*5*5*N)
## create molten data
mdf <- melt(arr, formula = . ~ p1 + p2 + p3 + N) # create molten data frame
## confidence interval calculated by `boxplot.stats`
f <- function(x){
ans <- boxplot.stats(x)
data.frame(x=x, y=ans$stats[3], ymin=ans$conf[1], ymax=ans$conf[2])
}
## (my poor) trial
ggplot(mdf, aes(x=p3, y=value)) + geom_boxplot(outlier.shape=1) +
stat_summary(fun.data=f, geom="rectangle", colour=NA, fill="black",
xmin=x-0.36, xmax=x+0.36, ymin=max(y-0.2, ymin), ymax=min(y+0.2,
ymax)) + facet_grid(p2 ~ p1, scales = "free_y")
**SOLUTION** (after the discussion with Kohske below):
f <- function(x, height){
ans <- median(x)
data.frame(y=ans, ymin=ans-height/2, ymax=ans+height/2)
}
p <- ggplot(mdf, aes(x=p3, y=value)) + geom_boxplot(outlier.shape=1) +
stat_summary(fun.data=f, geom="crossbar", height=0.5, colour=NA,
fill="black", width=0.78) +
facet_grid(p2 ~ p1, scales = "free_y")
pdf()
print(p)
dev.off()
**UPDATE** Hmmm... it's not that trivial. The following example shows that the "height" of the crossbar should be adapted to the y-axis scale, otherwise it might be overseen.
require(ggplot2)
require(reshape2)
require(plyr)
set.seed(1)
## parameters
p1 <- c(5, 20, 100)
p2 <- c("f1", "f2", "f3", "f4", "f5")
p3 <- c("g1","g2","g3","g4","g5")
N <- 1000
## lengths
l1 <- length(p1)
l2 <- length(p2)
l3 <- length(p3)
## build result array containing the measurements
arr <- array(rep(NA, l1*l2*l3*N), dim=c(l1, l2, l3, N),
dimnames=list(
p1=p1,
p2=p2,
p3=p3,
N=1:N))
for(i in 1:l1){
for(j in 1:l2){
for(k in 1:l3){
arr[i,j,k,] <- i+j^4+k+runif(N, min=-4, max=4)
}
}
}
arr <- arr + rexp(3*5*5*N)
arr[1,2,5,] <- arr[1,2,5,]+30
arr[1,5,3,] <- arr[1,5,3,]+100
## create molten data
mdf <- melt(arr, formula = . ~ p1 + p2 + p3 + N) # create molten data frame
f <- function(x, height){
ans <- median(x)
data.frame(y=ans, ymin=ans-height/2, ymax=ans+height/2)
}
## plot
p <- ggplot(mdf, aes(x=p3, y=value)) + geom_boxplot(outlier.shape=1) +
stat_summary(fun.data=f, geom="crossbar", height=0.7, colour=NA,
fill="black", width=0.78) +
facet_grid(p2 ~ p1, scales = "free_y")
pdf()
print(p)
dev.off()
here is an example:
f <- function(x, height) {
ans <- median(x)
data.frame(ymin = ans-height/2, ymax = ans+height/2, y = ans)
}
df <- data.frame(x=gl(2,6), y=c(1,1,1,1,3,3, 1,1,3,3,3,3))
ggplot(df, aes(x, y)) + geom_boxplot() +
stat_summary(fun.data = f, geom = "crossbar", height = 0.1,
colour = NA, fill = "skyblue", width = 0.8, alpha = 0.5)
if you just want to change the apparence, then here is a quick hack, I don't recommend though,
df <- data.frame(x=gl(2,6), y=c(c(1,1,1,1,3,3), c(1,1,3,3,3,3)*10))
ggplot(df, aes(x, y)) + geom_boxplot() + facet_grid(x~.)
gs <- grid.gget("geom_boxplot", grep = T)
if (inherits(gs, "grob")) gs <- list(gs)
gss <- llply(gs, function(g) g$children[[length(g$children)]])
l_ply(gss, function(g) grid.edit(g$name, grep=T, just = c("left", "center"), height = unit(0.05, "native"), gp = gpar(fill = "skyblue", alpha = 0.5, col = NA)))

Centering title in R viewport with multiple graphs ggplot2

I have a viewport with two graphs drawn in a left to right fashion. I'm trying to have a single title appear over both of the graphs centered in the viewport. Is this possible? Sorry, no pic because of my noob status and I can't post the code because of my work environment.
I think you might be interested in the gridExtra package, which provides the grid.arrange() function that fulfills everything you wonder about.
With #Kevin's example, the command would be
grid.arrange(plots[[1]], plots[[2]], ncol=2,
main="test main", sub="subtitle test")
I have been using an almost unmodified version of code in this post to the ggplot2 mailing list to put a main title and subtitle above and below a matrix of plots. Credit to Baptiste Auguié.
arrange <- function(..., nrow=NULL, ncol=NULL, as.table=FALSE,
main=NULL, sub=NULL, plot=TRUE) {
dots <- list(...)
n <- length(dots)
if(is.null(nrow) & is.null(ncol)) { nrow = floor(n/2) ; ncol = ceiling(n/nrow)}
if(is.null(nrow)) { nrow = ceiling(n/ncol)}
if(is.null(ncol)) { ncol = ceiling(n/nrow)}
fg <- frameGrob(layout=grid.layout(nrow,ncol))
ii.p <- 1
for(ii.row in seq(1, nrow)){
ii.table.row <- ii.row
if(as.table) {ii.table.row <- nrow - ii.table.row + 1}
for(ii.col in seq(1, ncol)){
ii.table <- ii.p
if(ii.p > n) break
fg <- placeGrob(fg, ggplotGrob(dots[[ii.table]]),
row=ii.table.row, col=ii.col)
ii.p <- ii.p + 1
}
}
if(!is.null(main) | !is.null(sub)){
g <- frameGrob() # large frame to place title(s) and content
g <- packGrob(g, fg)
if (!is.null(main))
g <- packGrob(g, textGrob(main), side="top")
if (!is.null(sub))
g <- packGrob(g, textGrob(sub), side="bottom")
} else {
g <- fg
}
if(plot) grid.draw(g)
invisible(g)
}
library(ggplot2)
plots <- llply(1:2, function(.x) qplot(1:10,rnorm(10), main=paste("plot",.x)))
arrange(plots[[1]],plots[[2]], nrow=1, ncol = 2, as.table=TRUE, main="test main",
sub="subtitle test")
Produces:
Another Possibility, it's kind of cheating by way of creating a blank plot...
p1 <- ggplot(diamonds, aes(price,depth)) + geom_point()
p2 <- ggplot(diamonds, aes(price,carat)) + geom_point()
p3 <- ggplot(diamonds, aes(x=1,y=1,label="Title")) + geom_text(size=20) + opts(panel.background=theme_blank(), panel.grid.minor=theme_blank(), panel.grid.major=theme_blank(), axis.text.x=theme_blank(), axis.text.y=theme_blank(), axis.ticks=theme_blank(), axis.title.x=theme_blank(), axis.title.y=theme_blank())
vplayout <- function(x, y) viewport(layout.pos.row=x, layout.pos.col=y)
grid.newpage()
pushViewport(viewport(layout=grid.layout(4,4)))
print(p1,vp=vplayout(2:4,1:2))
print(p2,vp=vplayout(2:4,3:4))
print(p3,vp=vplayout(1,1:4))

Resources