Data storytelling · Interactive practice

Bad Chart Clinic: diagnose it, fix it, get your skill profile

Fifty flawed charts and dashboards. For each one you spot what is wrong, pick the redesign you would ship, then see the real fix. At the end you get a skill radar and a level, from Newbie to Master.

50 cases8 skill axes5 levelsAll data is fictional · Runs offline in your browser

1. Diagnose

Tick every real flaw. Some options are decoys that sound plausible but are not problems; wrong ticks cost points, and ticking everything scores zero.

2. Choose the fix

Pick the redesign you would ship from four previews. One is best, others are partial or make things worse.

3. See the answer

Answers lock. You then see Before → After, the reason for every flaw, and the source page.

Choose a run

How scoring works

Teaching heuristic, not an exam. Each case is worth 100 points: 60 for diagnosis (each real flaw spotted earns its share; each decoy ticked costs 1.5 shares; select-all scores zero) and 40 for the fix (best 40, partial 10 to 25, harmful 0). Your radar uses accuracy per skill axis, shown with the number of items behind it. Axes with fewer than four items are marked low confidence. Your level depends on the average score and on how even your axes are.

The eight skill axes

Context & messageAudience, decision, Big Idea
Choosing the visualRight chart for the question
DeclutteringRemove what adds no information
Focus & colourGrey base, one accent
Text & titlesAction titles, labels, notes
Data integrityAxes, scales, honest geometry
AccessibilityColour-blind safe, legible, mobile
Dashboard designHierarchy, KPIs, filters, flow

Five levels

Inspired by the ideas in Storytelling with Data (Cole Nussbaumer Knaflic, 2015) and The Big Book of Dashboards (Wexler, Shaffer, Cotgreave, 2017). All cases, data and explanations are original and fictional. Page references point to the PDF page of the relevant discussion. This project is not affiliated with or endorsed by the authors or publishers. It is a learning aid built on our own reading of these ideas, so we may have misread or oversimplified them, and some references may be off by a page or two. Good design always depends on context, so treat the “right answers” as strong defaults, not absolute rules. Please read the books for the full treatment.