In jr-packages/jrGgplot2: Jumping Rivers: Advanced Graphics in R

Perhaps the easiest stat_* to consider is the stat_summary() function. This function summarises y values at every unique x value. This is quite handy, for example, when adding single points that summarise the data or adding error bars.

library("ggplot2")
data(bond, package = "jrGgplot2")
ggplot(bond, aes(Actor, Alcohol_Units)) +
             stat_summary(geom = "point", fun.y = mean)

A simple plot to create, is the mean alcohol consumption per actor (figure 1)

library("ggplot2")
data(bond, package = "jrGgplot2")
ggplot(bond, aes(Actor, Alcohol_Units)) +
  stat_summary(geom = "point", 
               fun.y = mean)

In the above piece of code we calculate the mean number of alcohol units consumed by each Actor. These x-y values are passed to the point geom. We can use any function for fun.y provided it takes in a vector and returns a single point. For example, we could calculate the range of values, as in figure 2:

ggplot(bond, aes(Actor, Alcohol_Units)) +
    stat_summary(geom = "point",
                 fun.y =  function(i) max(i) - min(i))

ggplot(bond, aes(Actor, Alcohol_Units)) +
             stat_summary(geom = "point", fun.y = mean)

Or we could work out confidence intervals for the mean number of Units consumed (figure 13):

## Standard error function
std_err = function(i)
    dt(0.975, length(i) - 1) * sd(i) / sqrt(length(i))

ggplot(bond, aes(x = Actor, y = Alcohol_Units)) +
    stat_summary(fun.ymin = function(i) mean(i) - std_err(i),
                 fun.ymax = function(i) mean(i) + std_err(i),
                 colour = "steelblue", geom = "errorbar",
                 width = 0.2, lwd = 2) +
    ylim(c(0, 20))

ggplot(bond, aes(x = Actor, y = Alcohol_Units)) +
    stat_summary(fun.ymin = function(i) mean(i) - std_err(i),
                 fun.ymax = function(i) mean(i) + std_err(i),
                 colour = "steelblue", geom = "errorbar",
                 width = 0.2, lwd = 2) +
    ylim(c(0, 20))

To calculate the bounds, we work out the standard deviation (sd(i)), then number of movies per actor (length(i)) and the correct value from the $t$ distributions, with $n-1$ degrees of freedom.

jr-packages/jrGgplot2 documentation built on Sept. 20, 2020, 2:59 a.m.