Module 3 — Data Visualisation & Storytelling

Edward Tufte, the godfather of data visualisation, established principles that remain essential: maximise the data-ink ratio (every drop of ink should present data, not decoration), avoid chartjunk (unnecessary gridlines, 3D effects, decorative elements), and tell the truth (never distort scales or cherry-pick data to mislead).

The purpose of visualisation is to make data understandable, actionable, and memorable. A good visualisation reveals patterns, trends, and outliers that would be invisible in a spreadsheet of numbers.

Key principles include: choose the right chart type for your data and message, label clearly (title, axes, units, source), use colour intentionally (to highlight, not to decorate), maintain consistent scales, and design for your audience (executives need different visualisations than analysts).

Chart selection should be driven by what you want to show:

- Comparison: Bar charts (categorical comparison), grouped/stacked bars (comparing across categories) - Trend over time: Line charts (continuous data), area charts (volume over time) - Part of whole: Pie charts (only for 2-5 categories), stacked bar charts, treemaps - Relationship: Scatter plots (correlation between two variables) - Distribution: Histograms (frequency distribution), box plots (statistical summary) - Geographical: Maps, choropleth maps

Dashboard design follows the principle of progressive disclosure: the most important information should be immediately visible, with the ability to drill down into details on demand. A well-designed dashboard answers the viewer's most pressing questions at a glance.

Best practices: limit to 5-7 visualisations per dashboard, use a consistent colour palette, align elements on a grid, place the most important information in the top-left (where eyes naturally start), and include context (benchmarks, targets, previous period comparisons).

Data storytelling combines data, visuals, and narrative to drive action. A data story follows a structure: setup (context and stakes), conflict (what the data reveals), and resolution (recommended actions).

The narrative structure for a data presentation might follow this pattern:

1. Hook: Start with a surprising finding or compelling question 2. Context: Establish what the audience needs to know about the situation 3. Findings: Present the data with clear visualisations 4. Implications: Explain what the findings mean for the business 5. Recommendations: Propose specific, data-supported actions 6. Call to action: What should happen next?

The biggest mistake in data communication is presenting analysis without a clear "so what?" Every chart, every number should serve the narrative and lead to a decision. If a data point doesn't help the audience decide or act, remove it.