Distributions

The shape of how a variable's values are spread: normal, right-skewed, left-skewed or uniform, with different modes and tails. Two products can share a mean and behave completely differently.

When to use

Check the shape before choosing a summary: for right-skewed data such as revenue, income or order values, report the median because the mean is misleading (df["col"].skew() is positive for a right tail). The shape is usually read from a Histograms chart.

See DS Lec 2.

Appears in

No references yet.