Add sequential arrows to a ggplot bubbleplot

Question

I have a line plot that I have placed an overlay of bubbles onto. Before I overlaid the bubbles, I was able to connect each point with an arrow to show the sequential relationship using this advice

But now that I have overlaid my bubbles, I still want to connect each bubble by an arrow as previous.

Here is my data:

X <- c(-0.373,-0.256,-0.272,0.048,0.219,0.313,0.209,0.112)
Y <- c(-0.055,-0.091,0.100,0.153,-0.139,-0.004,0.040,-0.004)
Size <- c(37,31,25,10,5,4,6,10)
Label <- c(1,2,3,4,5,6,7,8)

DF <- data.frame(X,Y,Size,Label)

Using the above advice, I can try and draw a plot with arrows connecting each bubble, but the size of the bubbles obscures the arrow heads.

ggplot(DF,aes(x=X, y=Y, size=Size,label=Label),legend=FALSE) + 
  geom_segment(aes(xend=c(tail(X,n=-1),NA), yend=c(tail(Y,n=-1),NA)), 
                   size=0.3, arrow=arrow(length=unit(0.3,'cm'))) + 
  geom_point(color='darkblue',fill="red", shape=21) +
  geom_text(size=2) +
  theme_bw() +
  scale_size(range = c(4, 30), name="Size", breaks=c(10,25,50),
             limits = c(1, 100))

I would basically like the above plot, but with the arrow heads visible. I know it is possible to write the arrows overtop the bubbles so I can see each arrow, but that is not what I am looking for. What I would like would be an arrow drawn from the outer edge of one bubble to the outer edge of the next bubble. So I need someway to shorten the head of each arrow by the radius of bubble it is pointing to.

And I have no idea why I get the warning at the end

Removed 1 rows containing missing values (geom_segment).

Answer 1

You can start with the following:

Size_penalty <- 1000
X <- c(-0.373,-0.256,-0.272,0.048,0.219,0.313,0.209,0.112)
X_next <- c(X[-1], NA)
Y <- c(-0.055,-0.091,0.100,0.153,-0.139,-0.004,0.040,-0.004)
Y_next <- c(Y[-1], NA)
Arrow_length <- sqrt((X - X_next)^2 + (Y - Y_next)^2)
Size <- c(37,31,25,10,5,4,6,10)
Size_next <- c(Size[-1], NA)
X_begin <- X + Size / Size_penalty * (X_next - X) / Arrow_length 
Y_begin <- Y + Size / Size_penalty * (Y_next - Y) / Arrow_length 
X_end <- X_next + Size_next / Size_penalty * (X - X_next) / Arrow_length 
Y_end <- Y_next + Size_next / Size_penalty * (Y - Y_next) / Arrow_length 
Label <- c(1,2,3,4,5,6,7,8)
DF <- data.frame(X, Y, X_begin, Y_begin, X_end, Y_end, Size, Label)

ggplot(DF, aes(x=X, y=Y, size=Size, label=Label),legend=FALSE) + 
  geom_point(color='darkblue', fill="red", shape=21) +
  geom_segment(aes(x=X_begin, y=Y_begin, xend=X_end, yend=Y_end), 
               size=0.3, arrow=arrow(length=unit(0.3, 'cm'))) + 
  geom_text(size=4) +
  theme_bw() + 
  scale_size(range = c(4, 30), name="Size", breaks=c(10, 25, 50),
             limits = c(1, 60))

Here I use Size / Size_penalty as a proxy to bubble radius, which is obviously quite far from being elegant. But this is the best I can do, since there's a scale_size, so that conversion from size to radius is implicit. All that is left is to find a conversion function like

rad <- function(ggplot_size_after_scaling) {}