The central idea

Data storytelling combines data + visuals + narrative so that an audience can understand an insight, care about it and decide what to do next. A chart without a message is only a display; a story without sound evidence is only an assertion.

Build the story

1. Audience — what does this stakeholder already know, need and care about?
2. Objective — are we informing, persuading or prompting action?
3. Key message — state the single takeaway before choosing a chart.
4. Narrative arc — establish context, reveal the problem or change, then resolve with an implication or call to action.
5. Evidence — use a small number of defensible insights; the PDF recommends the “power of three” to maintain focus.
6. Visual form — choose the chart for the relationship: trend, comparison, distribution, composition or association.
7. Guidance — use annotations, callouts and sequencing to direct attention to what matters.
8. Iteration — test with the intended audience, collect feedback and refine.

Choosing the visual

• Line chart: trend over time.
• Bar chart: comparison among categories.
• Histogram: distribution/frequency.
• Scatter plot: relationship between variables.
• Pie/stacked bar: parts of a whole, only when the comparison remains clear.
• Animation can stage a complex change, but it should clarify the causal or temporal sequence—not decorate the chart.

A FIT5120 data-story frame

Context → evidence → human meaning → product implication → action

Example structure:

• “Users are abandoning step 3” (claim).
• Show the relevant completion-rate evidence and denominator (proof).
• Explain which user need is blocked (meaning).
• Connect it to a feature or design decision (implication).
• Recommend the next experiment and success measure (action).

Quality and ethics checks

• Is the source current and credible?
• Are the denominator, timeframe and uncertainty visible?
• Does the chart exaggerate through scale, cropping or selective categories?
• Are affected communities represented rather than reduced to a statistic?
• Is the visual accessible: readable labels, sufficient contrast and no colour-only encoding?
• Can the audience distinguish observation, interpretation and recommendation?

Measure success

Useful measures include comprehension, audience engagement, decisions made, action taken and change in behaviour. Use feedback and analytics to improve the next version.

Sources

• Mastering the Art of Data Storytelling — Moodle PDF
• Data storytelling with animated graphs and charts
• Telling Data Stories