Circular plot in ggplot2 with line segments connected in r

Question

I am trying to create a circular plot and am stuck at a point:

dat1 <- data.frame (xvar = 1:10, y = 6, ymin = 4, ymax = 4.5)

Using this data, I can produce a circular ribbon plot in ggplot2

require(ggplot2) 
 ggplot(dat1, aes(x=xvar, y=y)) +  geom_ribbon(aes(ymin=ymin, ymax=ymax),
  col = "blue", fill = "blue2") + ylim (c(0,6)) + coord_polar()

However I want more.

I want to fill the segment of the ribbon with different colors and labels using the following data.

 filld <- data.frame (start = c(1, 4, 6, 7.5, 8, 9), end = c(4, 6, 7.5, 8, 9, 10),
                         label = c("A", "B", "C", "A", "C", "D"))
  filld
##    start  end label
## 1   1.0  4.0     A
## 2   4.0  6.0     B
## 3   6.0  7.5     C
## 4   7.5  8.0     A
## 5   8.0  9.0     C
## 6   9.0 10.0     D

The ribbon will be filled with different color by label variable. For example, the segment A will start from 1 and end at 4. Then segment B will start and end at 6 and filled with different color. Segments with same label (such as A and C) will be connected by line.

The resulting plot will look like this:

Answer 1

Here is an example:

filld$p <- rowMeans(subset(filld, select = c(start, end)))
ggplot(filld, aes(xmin = start, xmax = end, ymin = 4, ymax = 5, fill = label)) + 
  geom_rect() + 
  geom_segment(data = subset(filld, label %in% label[duplicated(label)]),
               aes(x = p, y = 0, xend = p, yend = 4, colour = label),
               size = 2, show_guide = FALSE) +
  geom_text(aes(x = p, y = 4.5, label = label), colour = "white", size = 10) +
  coord_polar() + 
  scale_y_continuous(limits = c(0, 5))

Updated

I do not recommend but something like this:

filld <- data.frame (start = c(1, 4, 6, 7.5, 8, 9), end = c(4, 6, 7.5, 8, 9, 10),
                     label = c("A", "B", "C", "A", "C", "D"))
filld$p <- rowMeans(subset(filld, select = c(start, end)))
filld <- merge(filld, ddply(filld, .(label), summarize, p2 = mean(p)))

lnd <- subset(filld, label %in% label[duplicated(label)])
lnd <- ddply(lnd, .(label), function(x) {
  x <- seq(x$p[1], x$p[2], length = 100)
  y <- 4.5 + ((x - mean(x))^2 - (x[1]-mean(x))^2) / (x[1]-mean(x))^2 * 3 + sin(x*3*pi) * 0.1
  data.frame(x, y)
})


p <- ggplot(filld, aes(xmin = start, xmax = end, ymin = 4, ymax = 5, colour = label, fill = label)) + 
  geom_line(aes(x, y, xmin = NULL, ymin = NULL, xmax = NULL, ymax = NULL), data = lnd, size = 2) +
  geom_rect() + 
  geom_text(aes(x = p, y = 4.5, label = label), colour = "white", size = 10) +
  coord_polar() + 
  scale_y_continuous(limits = c(0, 5))
p

Perhaps, what you want is beyond the scope of ggplot2.

Answer 2

Take a look at ggbio. This package extends ggplot2 and the grammar of graphics to sequence data (bioinformatics) but they seems to solve your problem (for example see here). Taking a look at the source code for the package may direct you to a more generic solution.

Answer 3

Add the ymin etc to fill d

fd <- data.frame(filld, ymin = 4, y =6, ymax = 4.5)

Then use geom_rect with the column label as the fill and colour aesthetics

 ggplot(fd, aes(x=start,xmin = start, xmax = end, y=y)) +  
   geom_rect(aes(ymin=ymin, ymax=ymax, fill = label )) + 
   ylim (c(0,6)) + 
   coord_polar() 

Add the lines:

## calculate the midpoint for each segment
fd$xmid <- apply(fd[,1:2],1,mean)
## get the replicate id for the labels (what occurence is this)
library(plyr)
library(reshape2)
fd1 <- ddply(fd, .(label), mutate, id = 1:length(xmid))
## reshape to wide, subsetting only with more than one rep
.lines <- na.omit(dcast(fd1, label~id, value.var = 'xmid'))
## add a mid point between the mid points
.lines$mid <- apply(.lines[,2:3],1,mean)
## reshape to long, and add some y values 
ld <- data.frame(arrange(melt(.lines,value.name = 'x'), label, x), y = rep(c(4,3,4),2))

ggplot(fd) +  
  geom_rect(aes(x=start,xmin = start, xmax = end, y=y,ymin=ymin, ymax=ymax, fill = label )) + 
  ylim (c(0,6)) + 
  coord_polar() + geom_path(data = ld, aes(x=x,y=y,colour = label))

The lines are ugly, but there!