Data Analysis › Charts & Visualization
Line Chart
Showing how a measure moves over an ordered scale, usually time.
Also known as: line graph, line plot, trend chart
A line chart shows how a measure moves over an ordered scale, usually time. The line’s job is direction and shape: rising, flat, seasonal, one-off step.
Reading one
Read it in three passes. First the overall direction — the trend. Then any repeating pattern, which is seasonality if it tracks the calendar. Then the points that break the pattern, which are the ones needing an explanation. That last pass is where the analysis usually is, and it is the one people skip.
What makes one good
- One measure, or a handful. Name each line directly on the line rather than in a legend at the edge.
- Time on the horizontal axis, evenly spaced. A gap in time is a gap on the axis, not a point to be skipped over.
- A vertical axis that fits the movement. Unlike a bar chart, a line chart may start above zero, because the reader compares the height of a point against the gridline rather than the length of a mark — see axis truncation.
- Smooth only when the underlying measure is smooth. A moving average over a spiky daily series is a modelling choice, and it should be labelled as one.
The classic mistake
A line chart over categories with no order. Connecting regions A to B to C produces a shape that means nothing, because the horizontal axis has no order to preserve. Lines join ordered scales. If the axis is unordered, use a bar chart.
The second mistake is too many lines. Past roughly five, the plot becomes spaghetti and the reader tracks none of them. Split by category into a grid of small charts instead, each with its own panel: small multiples. And whatever you discover needs a note on the chart — chart annotation — or the reader will invent a cause.