Data Visualization Fundamentals for UI

The plain-language idea

Every chart exists to answer one question, and the right chart type follows directly from which question that is:

• Comparison between categories ("which region sold the most?") → a bar chart. Bars give you length, and length is the easiest visual quantity for people to compare accurately.
• Trend over time ("is signup growth accelerating?") → a line chart. A line's slope makes direction and rate of change immediately readable in a way a series of disconnected bars doesn't.
• Part-of-whole ("what share of revenue is each product line?") → a stacked bar (or a pie chart, but only with 2-4 slices).
• Correlation between two variables ("does session length predict conversion?") → a scatter plot, where each point's position on two axes does the work.

The mismatch that shows up constantly in real products is a pie chart with 7 or 8 slices. It's a genuinely common default — "show proportions" reads as "use a pie chart" to a lot of people who reach for it without checking the slice count first. But human vision is bad at comparing angles and areas precisely; a pie chart with two or three slices is fine, but past four or five, the slices blur into "several small wedges of similar-looking size" and the viewer can't actually rank them. A horizontal bar chart, sorted largest to smallest, answers the exact same "what share of the whole" question and stays readable at 15 categories, not just 3.

See it in a real thing: the same data, two charts

Imagine a dashboard showing weekly active users by feature: Search 34%, Dashboard 28%, Export 19%, Settings 11%, Other 8%. As a pie chart, a viewer has to visually estimate five wedge angles and mentally rank them — doable, but it takes real effort and people get the ordering of the middle slices wrong more often than you'd expect. As a horizontal bar chart sorted descending, the ranking is just... the reading order, top to bottom, and the lengths let you compare Search to Dashboard at a glance without doing any angle math.

This is the deeper idea underneath chart-type choice: pre-attentive attributes. Length, position along a shared axis, and color hue are things the visual system processes in under 200 milliseconds, before conscious "reading" kicks in — that's what makes a well-chosen chart scannable. Angle, area, and especially 3D depth are not pre-attentive in the same way; they require deliberate, careful comparison. A 3D pie chart is the worst case of this stacked on itself: it adds a fake depth axis that distorts the apparent size of whichever slice faces the viewer, on top of the angle-comparison problem pie charts already have. It looks more "designed," but it measurably hurts comprehension of the actual data — which is the entire point of the chart existing.

What people usually miss

Watch out

• The question comes first, the chart type second. Picking a chart type by what "looks good" for the data shape, rather than by what question a viewer needs answered, is how bar charts end up used for trends and line charts end up used for unordered categories.
• A pie chart with more than 4-5 slices should almost always become a sorted bar chart. This isn't a stylistic preference — it's a direct consequence of angle/area comparison being harder for people than length comparison along a shared baseline.
• Color should be reserved for meaning, not decoration. Using a distinct hue for every category in a 12-category chart burns through the limited number of colors people can reliably distinguish and tell apart from a legend; grouping into "the top few, plus Other" is often more honest and more readable than showing everything.
• 3D and unnecessary depth effects don't just look outdated — they actively distort the data, because a 3D pie chart's front-facing slices appear proportionally larger regardless of their actual value.

Why this matters in interviews

"How would you visualize this dataset?" tests whether a candidate reasons from the question the stakeholder is asking to a chart type, or reaches for whatever chart is easiest to build. Naming the pie-chart-with-8-slices anti-pattern specifically is a strong signal of real dashboard experience, not textbook knowledge.