3.1 The pie chart — useful, but easy to misuse
A pie shows parts of a whole. It works only when there are a few slices (six or fewer), the categories don't overlap, and you're not trying to show a trend. Always label each slice — colour alone fails colour-blind readers.
The eye is bad at comparing angles. The moment you want the audience to compare sizes precisely, a pie fights you — a bar chart wins because the eye compares aligned lengths easily.
🔎 See it for yourself
Which two e-wallets are closest in student market share? Try to judge it from the pie…
- Touch ’n Go — 38%
- GrabPay — 27%
- Boost — 18%
- ShopeePay — 12%
- Others — 5%
Hard to rank by eye, right? Now the same data as a bar chart.
Never dress a pie with 3D or shadows — they distort the slice sizes. And if you have more than six categories, a pie becomes confetti; switch to a bar chart.
3.2 The scatterplot — showing a relationship
A scatterplot places one variable on the x-axis and another on the y-axis, one dot per observation. It answers: are these two things related, and how? The pattern can be positive, negative, none, or non-linear.
Plot each student's weekly revision hours (x) against CGPA (y). If the cloud of dots rises left-to-right, more revision is associated with higher CGPA — a positive relationship. A shapeless cloud means no clear relationship. (Association, not proof of cause — remind students of that.)
Each dot is one student. The dots rise left-to-right — a positive relationship. (Dashed line = the general direction; association is not proof of cause.)
You want to show (a) the share of students by faculty, and (b) whether taller students also weigh more. Which is a pie and which is a scatterplot — and why?
Points charts show one value each. A pie works for a few parts of a whole (≤6, labelled, no trend) — but the eye can't compare angles, so use bars to rank. A scatterplot shows the relationship between two variables (positive, negative, none, non-linear) — association, not cause.