A frequency table records how many times each value occurs in a set of data. To build one you go through the data and keep a tally — a running count drawn as strokes grouped in fives (four upright strokes |||| crossed by a fifth). Counting in fives makes a long list quick to total. The frequency is simply that total: how many of each value there are.

Suppose a shop notes the shoe size of every customer one morning. The tally and frequency for each size come out like this:

Notice the frequencies 3 + 5 + 8 + 4 = 20 add up to the total number of items — every customer is counted exactly once, so the frequencies must sum to the size of the whole data set.