Summary Statistics

A few numbers that describe a column: centre (mean, median, mode) and spread (standard deviation, quartiles). df.describe() produces them for the numeric columns.

When to use

Choose the measure to fit the data: default to the median for skewed business metrics such as revenue or salary, since the mean is outlier-sensitive (the lecture's salary example has a mean of 213k but a median of 15k). The lecture warns that summary statistics alone are never enough, so always visualise too (Anscombe's quartet).

See DS Lec 2.

Appears in

No references yet.