Graphical Representations for One Categorical Variable

A bar chart displays the distribution of one categorical variable: one bar per category, height equal to the category's frequency or relative frequency. A pie chart carries the same information as slices of one whole, which is exactly why it demands categories that sum to 100% of the units. The load-bearing skill is comparing groups of different sizes: raw counts inherit the group sizes, so the fair comparison divides each count by its own group's total and compares the resulting shares, ideally on a relative frequency bar chart where every group is drawn to the same percentage scale.

Two moves wreck this topic. The first is reading the taller count bar as the stronger tendency when the groups differ in size: 300 bus riders out of 1200 is a weaker lean than 180 out of 400, whatever the bars look like. The second is treating a bar chart like a histogram: hunting for skew in bars whose order is arbitrary, reading gaps between bars as missing data, or putting a genuinely categorical variable, however numeric its labels look, on a number line it does not have.