Draw a circle with ggplot2

Question

Maybe it is a silly question, but I couldn't find the answer in the handbook of ggplot2 nor with "aunt" google...

How do I plot a circle with ggplot2 as an additional layer if I have a middle point and a diameter? Thanks for your help.

Bobbiebobbin

Aunt Google was more responsive to me. This might be of some help. – Misleading

Answer 1

A newer, better option leverages an extension package called ggforce that defines an explicity geom_circle.

But for posterity's sake, here's a simple circle function:

circleFun <- function(center = c(0,0),diameter = 1, npoints = 100){
    r = diameter / 2
    tt <- seq(0,2*pi,length.out = npoints)
    xx <- center[1] + r * cos(tt)
    yy <- center[2] + r * sin(tt)
    return(data.frame(x = xx, y = yy))
}

And a demonstration of it's use:

dat <- circleFun(c(1,-1),2.3,npoints = 100)
#geom_path will do open circles, geom_polygon will do filled circles
ggplot(dat,aes(x,y)) + geom_path()

Misleading

Thanks Joran, that was what I'm looking for... I'm just wondering, that you have to go with the function to realize that in ggplot. If I remember it correctly, plot have a build in function for that. But this plots just look way nicer ;-) – Bobbiebobbin

@Bobbiebobbin - Indeed, there is grid.circle, but to make that work will require some knowledge of the grid system. – Misleading

Answer 2

If the purpose is only to annotate a circle, you can simply use annotate with geometry "path". No need to create a data frame or function:

#g is your plot
#r, xc, yc are the radius and center coordinates

g<-g+annotate("path",
   x=xc+r*cos(seq(0,2*pi,length.out=100)),
   y=yc+r*sin(seq(0,2*pi,length.out=100)))

Unbuild

Answer 3

Hi the following code from ggplot2 Google group may be useful:

dat = data.frame(x=runif(1), y=runif(1))
ggplot() + scale_x_continuous(limits = c(0,1)) +
scale_y_continuous(limits = c(0,1))+
geom_point(aes(x=x, y=y), data=dat, size=50, shape=1, color="gold4")

Which Produces:

I hope it gets you started in hacking up custom examples for your purpose.

Ora

I could be wrong, but I don't think this method addresses the OP's question, which was how to draw a circle given a center and diameter. It will be very tough to get the correct diameter using the size aesthetic. – Misleading

Ya I agree, it was just a pointer in the right direction not a complete solution. – Ora

Thanks for that hint Neo_me, but the way Joran posted fits better for my needs. – Bobbiebobbin

Great to see this, this was more what I was looking for. Thanks – Arthurarthurian

The reason why shape=1 is needed is that the default shape for geom_point is shape 19 where the fill color and stroke color cannot be controlled separately (so fill=alpha("black",0) doesn't work), but with shape=1 it's possible to use different fill and stroke colors and the default fill color is transparent. – Dowel

Answer 4

with ggplot2 >= 0.9 you can also do

library(grid)
qplot(1:10, 1:10, geom="blank") +
  annotation_custom(grob=circleGrob(r=unit(1,"npc")), xmin=2, xmax=4, ymin=4, ymax=6)

Laconism

This changes size with the size of the canvas. – Ultrared

Answer 5

For posterity's sake here is a more flexible circle solution using annotate and geom_ribbon that supports fill, color, alpha, and size.

gg_circle <- function(r, xc, yc, color="black", fill=NA, ...) {
    x <- xc + r*cos(seq(0, pi, length.out=100))
    ymax <- yc + r*sin(seq(0, pi, length.out=100))
    ymin <- yc + r*sin(seq(0, -pi, length.out=100))
    annotate("ribbon", x=x, ymin=ymin, ymax=ymax, color=color, fill=fill, ...)
}
square <- ggplot(data.frame(x=0:1, y=0:1), aes(x=x, y=y))
square + gg_circle(r=0.25, xc=0.5, yc=0.5)
square + gg_circle(r=0.25, xc=0.5, yc=0.5, color="blue", fill="red", alpha=0.2)

Barbell

Answer 6

Also try this,

 ggplot() + geom_rect(aes(xmin=-1,ymin=-1,xmax=1,ymax=1), fill=NA) + coord_polar()

The point being, a circle in some coordinates system is often not a circle in others, unless you use geom_point. You might want to ensure an aspect ratio of 1 with cartesian coordinates.

Laconism

What I want to plot is an importance horizon on a scatter plot. For this jorans solution seems to be the best way. But thanks a lot for your hint as well. – Bobbiebobbin