foundation Estimated learning time: 3 h

1.19 Descriptive statistics and distribution shape

You can summarize a dataset without being misled by the mean.

Before:00. Orientation & SetupUnlocks:03. Data Handling & Analysis04. Classical AI — Agents, Search & Knowledge Representation

Descriptive statistics are the first honest look at any dataset: centre, spread, shape, outliers. Quantiles and robust statistics matter because real data is skewed and contaminated, and the mean is easily dragged by a few extreme rows. It opens the statistics sequence because everything later summarises data somehow. The habit it installs is never reporting a mean without a histogram beside it — bimodal and heavy-tailed data make a lone mean actively misleading.

Work through these

  • Central tendency and spread

    Where a dataset sits and how spread out it is, using measures that respond differently to unusual values. The mean and the median disagreeing is itself information.

  • Quantiles, IQR, boxplots

    Quantiles describe a distribution by position rather than by arithmetic, which makes them resistant to extreme values. The box plot is their standard picture.

  • Robust statistics and outlier sensitivity

    Robust statistics answer the same questions while resisting the influence of a few strange points. Knowing which of your summaries are fragile is what stops one bad row moving a conclusion.

    Introduction to Probability & Statistics (18.05) · Course
  • Reading a histogram and a QQ plot

    A histogram shows shape and a quantile-quantile plot shows whether a distributional assumption is reasonable. Both are quick and both catch problems that summary numbers hide.

Sign in to keep your progress.

Free resources

Links last checked 29 Aug 2026.

Stuck here?

Ask a mentor. A real person answers, and they can see exactly which topic you're on. Usually within a couple of working days.

Checking your session…

Topics shown in module order.