5 · Analytics & data science

16. Visualisation and Analytical Storytelling

Chart choice, perceptual accuracy, dashboards and communicating uncertainty.

8 min read · 3 MCQs

Match chart to question

Lines for trends over time, bars for comparing categories, scatter for relationships, histograms for distribution shape, heatmaps for dense matrices. Pie charts work only for a couple of slices; humans compare angles poorly.

Honest encoding

Start bar axes at zero, avoid dual axes that manufacture correlation, sort categories meaningfully, and label directly instead of forcing legend lookups. Choose colourblind-safe palettes and reserve saturated colour for the point you are making.

Dashboards and narrative

A dashboard answers a recurring question and should fit one screen with the headline metric first, comparison second and detail on demand. For a one-off analysis, lead with the conclusion, show the evidence, and state the uncertainty and what would change your mind.

Chapter quiz

3 questions · pass mark 75%
  1. 1. Truncating a bar chart's y-axis is misleading because…

  2. 2. Dual axes are risky because they…

  3. 3. A good dashboard leads with…

Answer every question to submit. Progress for da-16 is saved in this browser.