Data Visualization interview questions
Data Visualization questions a Data analyst / BI can be asked, from first job to senior. Open one to see the strongest answer and why it works.
You have twenty-four months of revenue for three products. What chart do you draw, and why?
The strongest answer: A line chart with the months along the horizontal axis and one line per product — lines are what the eye reads as a trend, and the periods stay evenly spaced
Why it works: Right — time on the horizontal axis, one line per series, and the gaps between lines stay readable.
- Time goes on the horizontal axis.
- Lines show trends.
- One line per series.
- Keep the periods evenly spaced.
A chart says “Revenue 4.2.” What is missing?
The strongest answer: Units, currency, tax, and the period — 4.2 of what, for whom, when
Why it works: Exactly. A number without units is a riddle. Arrows are not units.
- Units and currency.
- Tax in or out.
- The period.
- Do not leave 4.2 naked.
What must every dashboard show besides the chart?
The strongest answer: When the data was refreshed and which definition it uses — a pretty stale number is still stale
Why it works: Exactly. Freshness and definition. “Live” theatre is a red flag.
- Show refresh time.
- Show the definition.
- Do not badge nightly as live.
- Do not hide staleness.
A bar chart starts at 97 so a 1-point change looks huge. What do you do?
The strongest answer: Start at a zero or say why not — I do not crop the axis to manufacture drama
Why it works: Exactly. Axes are part of the claim. Cropping is a rhetorical trick you must own or drop.
- Zero-baseline or a written reason.
- Do not crop for drama.
- Keep the labels.
- Skip 3D tricks.
A dashboard opens with one huge number and nothing else: 48,210 active users.
What is missing from that tile?
The strongest answer: Everything that makes it mean something — the period it covers, a comparison against the previous period or the target, and a small trend line, so the reader can tell whether it is good news
Why it works: Exactly — a number alone cannot be read. Period, comparison and trend turn it into something someone can act on.
- A number needs a comparison.
- State the period covered.
- Show it against target or last period.
- Add the trend, not more digits.
A director wants the axis cropped and the outlier dropped “for the board.” What do you do?
The strongest answer: Offer a full-range view and a noted exception — I will not drop an outlier that is real money without a sentence on the slide
Why it works: Exactly. You can frame; you cannot un-happen money. Private copies and warehouse deletes are red flags.
- Show the range or note the drop.
- Real money gets a sentence.
- No second set of books.
- Do not delete warehouse rows for a slide.
Your chart is titled “Checkout completion by day.” What would a stronger title say?
The strongest answer: What I found — completion dropped after the 3 September release and has not recovered — with a subtitle giving the period covered and what the chart does not prove
Why it works: Exactly — the title is the one sentence everyone remembers, so it carries the finding and its limits.
- The title is the finding.
- One sentence, in plain words.
- Subtitle: period and caveat.
- Claim only what the chart shows.
A dashboard has 40 charts and nobody opens it. What do you cut toward?
The strongest answer: The weekly decision it was for — three numbers and a residual — I do not add a 41st chart to “increase engagement”
Why it works: Exactly. A dashboard is a task. Engagement theatre is how it dies louder.
- One weekly decision.
- Few numbers + residual.
- Do not add chart 41.
- Do not spam the 40.
A slide shows ad spend and revenue on one chart with two different vertical axes; the lines cross and then track each other closely.
What do you say about it?
The strongest answer: That the tracking is a choice of scales, not a finding — I would stack two charts sharing the time axis, or index both series to a common base, and read nothing into the crossing point
Why it works: Right — two free scales can make almost any pair of lines look related. Remove the illusion rather than annotate it.
- Two axes can fake a relationship.
- Stack charts on a shared time axis.
- Or index both to a base.
- Never read the crossing point.
Across your team’s reports the same product is blue in one chart and orange in the next, and red means both “bad” and “Europe.”
What do you standardise?
The strongest answer: A small palette with fixed meanings — one colour per series everywhere, grey for context, status colours reserved for status — checked against colour-blind readers, and never the only signal
Why it works: Right — consistency, meaning and a label. Colour that carries information has to survive a colour-blind reader and a printed page.
- Colour must mean one thing.
- Fix a series colour everywhere.
- Grey the context, accent the point.
- Never let colour be the only cue.
A beautiful dashboard changed no decision in six months. What do you do?
The strongest answer: Retire it or attach it to a forum that decides — beauty without a decision is a cost centre
Why it works: Exactly. Analytics is decisions. Wider mail and awards are how theatre grows.
- Attach a decision forum or retire.
- Beauty is not a KPI.
- Do not widen the spam.
- Do not award a unused deck.