Coders sometimes want to view not the raw counts of the observations in each bin (for example, heart rate was between 65-75 bpm on 5 days), but instead the proportion of the total data that are contained in that bin (for example, heart rate was between 65-75 bpm on 25% of the days).
So rather than having "number of observations in this bin" on the y axis, we would have "number of observations in this bin divided by total number of observations from all bins". (This is a simplification, but basically that's what the graph represents.)
This type of graph is called a density histogram.
Use gf_dhistogram() to create a density histogram of sleep.
Density plots are similar to density histograms, except there aren't really any bins. They show a smoother graph of essentially the same thing as a density histogram.
Sometimes it is useful for coders to view two representations of the same data at the same time. To use gf_density() and gf_dhistogram() together, use the conjunction %>%. This operator is called the pipe, and you can use it to simply stick one graph on top of another one.
Now it's much easier to see the underlying shape of the data.