Density over histogram using ggplot2

Question

I have "long" format data frame which contains two columns: first col - values, second col- sex [Male - 1/Female - 2]. I wrote some code to make a histogram of entire dataset (code below).

ggplot(kz6, aes(x = values)) + 
  geom_histogram()

However, I want also add a density over histogram to emphasize the difference between sexes i.e. I want to combine 3 plots: histogram for entire dataset, and 2 density plots for each sex. I tried to use some examples (one, two, three, four), but it still does not work. Code for density only works, while the combinations of hist + density does not.

density <- ggplot(kz6, aes(x = x, fill = factor(sex))) + 
  geom_density()

both <- ggplot(kz6, aes(x = values)) + 
  geom_histogram() +
  geom_density()

both_2 <- ggplot(kz6, aes(x = values)) + 
  geom_histogram() +
  geom_density(aes(x = kz6[kz6$sex == 1,]))

P.S. some examples contains y=..density.. what does it mean? How to interpret this?

Rui Barradas Rui Barradas · Accepted Answer · 2020-02-28T19:41:04

To plot a histogram and superimpose two densities, defined by a categorical variable, use appropriate aesthetics in the call to geom_density, like group or colour.

ggplot(kz6, aes(x = values)) +
  geom_histogram(aes(y = ..density..), bins = 20) +
  geom_density(aes(group = sex, colour = sex), adjust = 2)

Data creation code.

I will create a test data set from built-in data set iris.

kz6 <- iris[iris$Species != "virginica", 4:5]
kz6$sex <- "M"
kz6$sex[kz6$Species == "versicolor"] <- "F"
kz6$Species <- NULL
names(kz6)[1] <- "values"
head(kz6)

Density over histogram using ggplot2

1 Answers