Best Practices and Data Storytelling

The ultimate goal of data visualisation is not just to display data — it is to communicate a message. The most technically sophisticated chart fails if the audience does not understand it or does not care. This lesson covers best practices for chart design, common mistakes to avoid, and the art of data storytelling.

---

The Data Storytelling Framework

A data story has three components:

Component	Description
Data	The foundation — accurate, relevant, well-prepared
Narrative	The story arc — context, conflict, resolution
Visuals	The charts — chosen and designed to support the narrative

A good data story answers: What happened? Why does it matter? What should we do about it?

---

The Narrative Arc

Every compelling data story follows a structure:

1. Setup (Context)

Provide background. Why are we looking at this data? What is the business question?

2. Rising Action (Exploration)

Show the data. What patterns emerge? What is surprising or concerning?

3. Climax (Insight)

Deliver the key finding — the "aha moment" that changes understanding.

4. Resolution (Recommendation)

What should the audience do with this insight? What action follows?

---

Best Practices Checklist

Titles and Labels

Practice	Example
Use descriptive titles that state the insight, not just the topic	"Sales grew 23% in Q4" instead of "Q4 Sales"
Label axes with units	"Revenue (GBP thousands)"
Use direct labels on chart elements instead of a legend when possible	Place labels next to lines
Keep axis labels horizontal	Rotate only when absolutely necessary

Colour

Practice	Reason
Use colour purposefully	Every colour should encode meaning
Highlight the key data point, grey out the rest	Draws attention to the insight
Use accessible palettes	8% of men have colour vision deficiency
Be consistent	Same colour = same meaning across charts

Layout and Composition

Practice	Reason
Remove chart junk (3D, unnecessary gridlines, borders)	Increases data-ink ratio
Align charts for comparison	Side-by-side with shared axes
Use white space deliberately	Prevents visual overload
Order categories meaningfully (by value, not alphabetically)	Makes comparison easier

---

Common Mistakes to Avoid

Mistake	Why It's Harmful	Fix
Truncated y-axis	Exaggerates differences	Start bar charts at zero
Dual y-axes	Implies false correlation	Use two separate charts
Too many colours	Overwhelms the viewer	Limit to 5-7 colours
Pie chart with many slices	Impossible to compare angles	Use a bar chart
Missing context	Numbers without comparison are meaningless	Add benchmarks, targets, or prior periods
Decorative elements	Distract from the data	Simplify — less is more
Inconsistent scales	Misleading comparison across charts	Use shared axes and scales
Rainbow colour maps	Not perceptually uniform	Use sequential or diverging palettes

---

Highlighting Techniques

Grey-and-Highlight Strategy

Use grey for all data points except the one you want the audience to focus on.

import matplotlib.pyplot as plt

countries = ["UK", "France", "Germany", "Italy", "Spain"]
gdp = [3070, 2780, 4070, 2000, 1390]

colours = ["#CCCCCC"] * len(countries)
colours[2] = "#E74C3C"  # Highlight Germany

fig, ax = plt.subplots(figsize=(8, 5))
ax.barh(countries, gdp, color=colours)
ax.set_xlabel("GDP (Billion USD)")
ax.set_title("Germany leads European GDP")
ax.spines[["top", "right"]].set_visible(False)
plt.tight_layout()
plt.show()

Annotations

Draw attention to specific data points with text and arrows.

import matplotlib.pyplot as plt

months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun",
          "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"]
revenue = [42, 45, 48, 52, 38, 41, 55, 58, 63, 67, 71, 80]