Julia Watzek
Developing Data Products course
This application is a visualization for random samples drawn from a binomial distribution with the following user-specified parameters.
n)size)prob)Users don't need to repeatedly execute code and, in fact, don't need knowledge of R at all.
library(ggplot2)
n = 100; size = 15; prob = .5
df = data.frame(x = rbinom(n, size, prob))
( g = ggplot(df, aes(x = x)) + theme_bw() + scale_x_continuous(limits = c(0, size)) +
geom_histogram(aes(y = ..density..), binwidth = 1, colour = 'black', fill = 'grey95') )
The density of the normal distribution with mean and standard deviation of the generated sample can be superimposed on the plot.
g + stat_function(fun = dnorm, args = list(mean = mean(df$x), sd = sd(df$x)),
colour = 'steelblue', size = 1)
This makes the app a useful tool for statistics teachers to illustrate the normal approximation of the binomial distribution when n is large enough.
It also nicely demonstrates the concept of repeated sampling from a known probability distribution, which facilitates understanding of null-hypothesis significance testing.