Storytelling with data

Data storytelling connects analysis to action: context, conflict (the problem), insight (what data show), and recommendation (what to do next). Numbers support the narrative—they are not the whole story.

Narrative arc

1. Hook — why should the audience care now?
2. Context — metric definitions, time window, population
3. Insight — one main finding per section
4. Implication — risk or opportunity in business terms
5. Recommendation — specific next step with owner

Audience tailoring

• Executives — decision, impact, confidence; appendix for methods
• Engineers — data lineage, edge cases, reproducibility
• Analysts — SQL logic, assumptions, sensitivity checks

Anti-patterns

• Dumping 20 charts without a thesis
• Burying the recommendation in jargon
• Hiding limitations to look smart—erodes trust

Slide and notebook hygiene

One headline per slide stating the insight. In Jupyter, markdown cells explain why before code cells show how.

Important interview questions and answers

1. Q: Insight vs data?
   A: Insight is interpreted meaning; raw tables require audience to do your job.
2. Q: Recommendation test?
   A: Can someone act Monday morning from your conclusion without re-running analysis?

Self-check

1. List five parts of a data narrative arc.
2. How should executive summaries differ from engineer docs?
3. Name one storytelling anti-pattern.

Tip: Lead with the recommendation, then evidence.