SQL, data modeling, data quality, EDA, visualization · insight-0072
Index series to a common base period instead of using a dual-axis chart
- Evidence type
- practitioner heuristic
- Difficulty
- easy
- Roles
- Data analyst, Data scientist
- Industries
- aviation, retail, finance
Summary: situation, action, reason, boundary
Situation
You want to compare growth of revenue (in millions) with passengers (in thousands) on one chart.
Action
Divide each series by its value in a stated base period and multiply by 100, plot them on one axis, and label the base ("Jan 2023 = 100"). Choose a base period that is typical, or show robustness to the base.
Reason
Indexing puts different units on one relative scale, so divergence reflects different growth rather than arbitrary axis choices.
Boundary
Comparing growth of KPIs with different units, of segments of very different size, or of prices and volumes.
Symptom or risk
A dual-axis chart lets the axis ranges, not the data, decide whether the lines cross or diverge, so any story can be drawn.
Exceptions
When absolute levels matter (capacity versus demand in seats), plot levels on the same unit instead.
Trade-offs
Indexing hides levels; add them in a table or tooltip.
Pitfalls
Choosing an unusual base month such as a strike or holiday, which distorts every later value.