A box-and-whisker plot summarises one or many numeric columns, with whiskers reaching 1.5·IQR and anything beyond drawn as an outlier; adding a grouping column splits the first numeric by group.
Tukey's box plot summarises a distribution with five numbers: min (within 1.5·IQR), Q1, median, Q3, max (within 1.5·IQR). The whiskers stop at the last value within the fence, and anything past is plotted as an individual outlier dot — giving you a robust shape sketch + outlier flags in one chart.
Box plots are the right choice for comparing distributions across groups: any two groups can be visually contrasted by the relative position of their boxes (median shift), the overlap of their notches (heuristic significance), and the spread of their whiskers (variance differences). For showing distribution shape on a single variable, the violin plot is more informative.
Numeric columns; optional categorical grouping column.
Box per group / column with min / Q1 / median / Q3 / max + outliers.
If the median sits off-centre in the box, the underlying distribution is skewed.
Boxes of very different heights ⇒ unequal variances ⇒ favour Welch / Brown-Forsythe ANOVA over classical.
When you want to see distribution shape (bimodality, plateau), switch to a violin plot or histogram — the box hides it.