Creating arrow head matching size (or lwd) in ggplot2

Question

To draw arrows in ggplot, I use geom_segment and arrow=arrow(). I would like the arrow head size to match the segment width (or size). However, arrow does not recognize variables directly from the data argument in ggplot and one must specify data.frame containing the variable using the $ operator. This causes a disjunct between the values used for plotting the line and those used for plotting the arrow head (the largest arrow head can be on the thinest segment).

Example:

d <- structure(list(Process = structure(c(2L, 1L, 1L, 1L, 2L, 2L, 
1L, 1L, 2L, 1L, 2L), .Label = c("First", "Second"), class = "factor"), 
x.sink = c(1, 3, 1, 2, 2, 3, 3, 2, 2, 2, 2), y.sink = c(1, 
1, 1, 2, 2, 1, 1, 1, 1, 2, 2), x.source = c(2, 2, 2, 2, 2, 
2, 2, 1, 1, 1, 3), y.source = c(2, 2, 2, 1, 1, 1, 1, 1, 1, 
2, 1), offset = c(1, 1, 1, -1, -1, -1, -1, -1, -1, 1, -1), 
Std.Flux = c(0.179487179487179, 0.170940170940171, 0.944444444444444, 
0.0854700854700855, 0.726495726495726, 0.128205128205128, 
0.213675213675214, 0.213675213675214, 0.128205128205128, 
0.106837606837607, 1)), .Names = c("Process", "x.sink", "y.sink", 
"x.source", "y.source", "offset", "Std.Flux"), class = "data.frame", row.names = c(NA, 
-11L))


p <- qplot(data=d,
           #alpha=I(0.4),
           colour=Process,
           size=Std.Flux,
           xlim=c(0,4),
           ylim=c(0,3),
           x=x.source+as.numeric(Process)/10,
           y=y.source+as.numeric(Process)/10,
           xend=x.sink+as.numeric(Process)/10,
           yend=y.sink+as.numeric(Process)/10,
           geom="segment",
           arrow = arrow(type="closed",
                         length = unit(d$Std.Flux,"cm")))
print(p)

Disjunct arrow heads

Any suggestions?

i reckon it's an oversight; worth posting a feature request on github — baptiste
Voting to close - Reporting bug which must have been fixed in the last 8 years - not reproducible any more (see reprex in my answer) — tjebo

www www · Accepted Answer · 2017-09-17T05:23:48

Here's one way:

require(ggplot2)

df <- mtcars

arrow_pos <- data.frame(y = 250)

ggplot(df, aes(x=factor(cyl), y=mpg)) +
  geom_bar(width = .4, stat="identity", fill="darkblue")  +
  geom_segment(data=arrow_pos, 
               aes(x=1.526, xend=1.01, y=y + 90.02, yend=y + 0.25),
               arrow=arrow(length=unit(4.2, "mm")), lwd=2,
               color="black") +
  geom_segment(data=arrow_pos, 
               aes(x=1.525, xend=1.01, y=y + 90, yend=y + 0.25),
               arrow=arrow(length=unit(4, "mm")), lwd=1,
               color="gold2") +
  annotate("text", x=2.39, y=360, 
           label='This arrow points to the highest MPG.') +
  scale_y_continuous(limits = c(0,400)) +
  xlab('CYL') + ylab('MPG')

Output:

Creating arrow head matching size (or lwd) in ggplot2

2 Answers