For many large data sets, there is a clear pattern to the distribution.

Statisticians have found ways to produce formulas whose graphs fit a great variety of shapes, including this one.

When superimposed, we see that the curve does a pretty good job of matching the histogram.

Since area in histograms is proportional to the number of values, if we arrange the units on the vertical axis so that the total area is 1 ( making it a density curve ) , then the area under the curve will be the proportion of the total number of values: