Equivalent of curve() for ggplot

问题:

Is there an equivalent method for plotting functions using ggplot to the curve() command employed in base graphics? I guess that the alternative would be to just create a vector of values of the function and plot a connected line, but I was hoping for something a bit more simple.

Thanks!

回答1:

You can add a curve using the stat_function:

ggplot(data.frame(x=c(0, 10)), aes(x)) + stat_function(fun=sin)

If your curve function is more complicated, then use a lambda function. For example,

ggplot(data.frame(x=c(0, 10)), aes(x)) + 
  stat_function(fun=function(x) sin(x) + log(x))

you can find other examples at http://kohske.wordpress.com/2010/12/25/draw-function-without-data-in-ggplot2/

---

In earlier versions, you could use qplot, as below, but this is now deprecated.

qplot(c(0,2), fun=sin, stat="function", geom="line")

回答2:

The data.frame example above works well, and it makes grid lines. The qplot example doesn't work in ggplot2 2.2.0 for the reasons given.

You can also use the "curve" function in ggplot2 2.2.0, but it does not automatically make grid lines or background color. For example:

curve(cos(x), from= 0, to= pi/2).  

The "ggplot(data.frame(... ) method gives the full impressive range of ggplot2's formatting options. I like it.