In dtkaplan/MultipleChoice: Support for Multiple Choice Questions in Shiny

require(statisticalModeling)
require(mosaic)
require(dplyr)

I'M NOT SURE WHAT TO DO WITH THIS ONE. IT COULD GO IN THE CHAPTER "What is explaining an outcome?" The distribution of the groups are narrower than the distribution of the whole.

Sometimes when making a graphic, you want to change some aspect of it or add new elements to it. To illustrate, here are commands for making simple density plot in the ordinary way.

gf_density( ~ height + fill:"black", data = Galton) %>%
gf_density( ~ height + fill:sex, alpha = 0.25, data = Galton)

ggplot() + 
  geom_density(data = select(Galton, height), aes(x = height), fill = "gray") +
  geom_density(data = Galton, aes(x = height, fill = sex, color = sex), alpha = .5) +
  facet_wrap( ~ sex)

In looking at the plot, you realize that you want a better label for the horizontal axis. You can do this by adding on a new component to the plot specifying the axis label.

gf_density( ~ height, data = Galton) + xlab("Child's height (in)")

Use the + sign to add components that don't start with gf_ such as xlab(), ylab(), facet_wrap(), and so on. Now suppose you decide to superimpose density plots of the mothers' and fathers' heights. To plot each of these variables individually, you would use a command like this:

gf_density( ~ mother, data = Galton)

But we want to overlay the distribution of both the fathers and mothers so we can readily compare them to the children. For this purpose, a bit of data wrangling is required: creating one variable with the parents' heights and another indicating the sex of the parent.

Parents <- 
  Galton %>%
  group_by(family) %>%
  filter(row_number() == 1) %>%
  ungroup() %>%
  select(father, mother) %>% 
  tidyr::gather(key = parent, value = parent_height, father, mother)

You can add this component to the graph by connecting it to the previous components. Use %>% to add a component that does start with gf_.

gf_density( ~ height + fill:sex, alpha = 0.5, color = NULL, data = Galton) %>%  
  gf_density( ~ parent_height + fill:parent, alpha = 0.3, color = NULL, data = Parents) +
  xlab("Child's height (in)") + 
  facet_wrap( ~ sex)