R polygon: fill values >0 in a density plot

664 views Asked by At

I have the following:

set.seed(513)
x1 = rnorm(12^6, 5,5)
x2 = x1-10
plot(density(x1),xlim = c(-25,25), ylim = c(0,0.09), xaxs = "i", yaxs = "i", main ="", xaxt = "n", yaxt = "n", xlab = "", ylab = "", col = "blue")
lines(density(x2), col = "red")
axis(side = 1, pos = 0, at = c(-30,-20,-10,0,10,20,30), labels = c("",-2,-1,0,+1,+2,""), tck = -0.02)
lines(x=c(0,0),y=c(0,1), lty = 3)
polygon(density(x1[x1>=0]), col = rgb(0,0,0.5,0.5))
polygon(density(x2[(x2)>=0]), col = rgb(0.5,0,0,0.5))
par(xpd=TRUE)
text("Frequency", x = -29, y = 0.045, srt = 90)
par(xpd=FALSE)

I would like to make a plot which shows two normal curves, with the space where X>0 having filled colours. I tried using polygon but it seems to struggle with density (it seems using x1[x1>=0] limits the density function to that data, I think I need it to make the distribution first and then cut it)

I would like to fill the space under the curves for both lines where X is zero or larger.

enter image description here

1

There are 1 answers

0
rg255 On BEST ANSWER

I managed to make a rather inelegant solution...

set.seed(513)
x1 = rnorm(12^6, 5,5)
x2 = x1-10
plot(density(x1),xlim = c(-25,25), ylim = c(0,0.09), xaxs = "i", yaxs = "i", main ="", xaxt = "n", yaxt = "n", xlab = "", ylab = "", col = "blue")
lines(density(x2), col = "red")
axis(side = 1, pos = 0, at = c(-30,-20,-10,0,10,20,30), labels = c("",-2,-1,0,+1,+2,""), tck = -0.02)
lines(x=c(0,0),y=c(0,1), lty = 3)

x = density(x1)[1]
y = density(x1)[2]
x = as.numeric(x[[1]])
y = as.numeric(y[[1]])
df = data.frame(x,y)
polygon(c(df$x[df$x>=0],0),c(df$y[df$x>=0],0),col = rgb(0,0,0.5,0.5))
x = density(x2)[1]
y = density(x2)[2]
x = as.numeric(x[[1]])
y = as.numeric(y[[1]])
df = data.frame(x,y)
polygon(c(df$x[df$x>=0],0),c(df$y[df$x>=0],0),col = rgb(0.5,0,0,0.5))

par(xpd=TRUE)
text("Frequency", x = -29, y = 0.045, srt = 90)
par(xpd=FALSE)

enter image description here