Showing Trends and Comparisons Honestly

Lesson 5 of 8 · Data Visualization Fundamentals · Beginner

Charts that show trends and comparisons can inform good decisions — or mislead audiences into drawing incorrect conclusions. Honest visualization requires both accurate values and appropriate framing.

Time Series and Baselines

A time series shows how a value changes over time. The baseline — where the vertical axis starts — has a significant effect on how change appears. For bar charts, the axis should generally start at zero. A bar chart with a truncated axis makes small differences look large, which can mislead readers about the magnitude of change.

Line charts may use a non-zero baseline when the range of variation is the focus and the baseline is clearly labelled. If a line chart shows values between 980 and 1,020, starting the axis at zero would compress the variation into a nearly flat line. In this case, a non-zero baseline with a clear label is appropriate — but the choice should be documented and the axis clearly marked.

Comparable Scales

When comparing two or more series, use the same scale for all. Dual-axis charts — where two series use different vertical axes — can make unrelated trends appear correlated. Use dual axes only when the relationship between the two series is the explicit subject of the chart, and label both axes clearly.

Absolute Values, Percentages, and Rates

Choosing between absolute values, percentages, and rates changes what the chart communicates:

• Absolute values show the raw count or amount. Useful when the total matters.
• Percentages show proportion. Useful for comparing groups of different sizes.
• Percentage points measure the arithmetic difference between two percentages. A change from 20% to 25% is a 5 percentage-point increase — not a 5% increase. Confusing these two is a common and consequential error.
• Rates (such as per 1,000 residents) normalize for population size, making comparisons across areas of different sizes more meaningful.

Missing Periods and Irregular Intervals

A time series with missing months or irregular intervals can create a misleading impression of continuity. If data is missing for certain periods, note this clearly — either in the chart annotation or in accompanying text. Do not connect data points across a gap without indicating that the gap exists.

Uncertainty and Provisional Data

Data is often provisional, estimated, or subject to revision. Show uncertainty where it exists — through annotations, notes, or visual indicators. Presenting provisional data as final, or omitting the fact that figures may be revised, misleads readers about the reliability of the information.

Selective Date Ranges

Choosing a start or end date that makes a trend look more dramatic than the full picture supports is a form of misleading framing. Show the full available time range, or clearly explain why a specific range was chosen.

Practical Example

A community centre reports that program attendance "increased by 25%" over the past year. The full picture: attendance rose from 400 to 500 visits per month — a 100-visit increase, or 25% relative to the prior year. The previous year had unusually low attendance due to a temporary closure. Showing only the most recent year-over-year comparison without this context overstates the trend. A chart showing three years of monthly data with an annotation explaining the closure gives a more honest picture.

Common Mistakes

• Truncating bar-chart axes to exaggerate differences.
• Confusing percentage change with percentage-point change.
• Using dual axes to imply a relationship between unrelated series.
• Selecting a date range that makes a trend look more significant than the full data supports.
• Presenting provisional data as final without noting that figures may change.

Key Takeaways

• Bar charts should generally start at zero; line charts may use a non-zero baseline when clearly labelled.
• Percentage change and percentage-point change are different measures — specify which you are using.
• Use rates or percentages when comparing groups of different sizes.
• Note missing periods and provisional data clearly.
• Show the full available time range, or explain why a specific range was chosen.