Skip to content
ML Academy

SQL, data modeling, data quality, EDA, visualization · insight-0068

Choose the chart from the comparison the reader must make, not from the data type

Evidence type
practitioner heuristic
Difficulty
easy
Roles
Data analyst, Data scientist
Industries
aviation, cross-industry

Summary: situation, action, reason, boundary

  1. Situation

    You need to show revenue share by fare level across three markets and the default tool suggests three pie charts.

  2. Action

    Write the comparison in one sentence ("compare high-fare share across markets"), then pick the encoding that puts those values on a common scale: side-by-side bars for comparing parts, stacked bars for parts of a whole across conditions, a line for change over time. Use pies only for two or three simple parts.

  3. Reason

    Charts are read most accurately when the compared values share an aligned axis; pies and stacked segments make readers compare angles or floating lengths.

  4. Boundary

    Reports, dashboards and slides.

Symptom or risk

Readers cannot compare slice sizes across pies, so the key message (one market relies on high fares) is lost.

Exceptions

A single part-to-whole message with few parts can work well as a pie or a stacked bar.

Trade-offs

The most accurate chart may look less familiar to some audiences.

Pitfalls

Adding more series to one chart instead of choosing which comparison matters.

Lessons and concepts

Claims and their sources

  • Pie charts work best for simple fractions and small datasets, stacked bars suit comparisons across conditions but make non-baseline segments hard to compare, and side-by-side bars are best for comparing individual fractions directly.

    Fundamentals of Data Visualization