This entry is part of the Chart Index, the reference library for the Chart Design Field Guide.
The line chart is the natural form of change over time. Where a bar chart asks the eye to compare lengths, a line chart asks it to follow a trajectory. The movement becomes the message: slope, acceleration, seasonality, volatility, plateau, recovery.
That fluency is also the risk. A line implies continuity. If the x-axis is not ordered, or if values are missing, the connecting stroke can invent a relationship that is not in the data.
What it is
A line chart maps an ordered or continuous variable, usually time, to one axis and a quantitative value to the other. Each observation is plotted as a point; the points are connected so the reader can see direction and rate of change.
The hero chart above is a typical line-chart problem: three related series, one shared time axis, and a question about trajectory. The reader can see that all three platforms grew, but mobile changed fastest and the partner channel barely moved.
When to use it
Line charts are the right choice when:
- The x-axis is ordered or continuous: time, distance, temperature, dose, age, or another sequence where neighbouring values are meaningful.
- The reader's question is "how is this changing?" rather than "which category is bigger?"
- You have enough observations to see a pattern. Seven to ten points can show direction; daily or weekly data often needs many more.
- You want to compare two to five related series on the same scale.
- The rate of change matters. Slope is the visual encoding, and slope is what the reader will remember.
When not to use it
- Discrete categories. If the x-axis is regions, products, teams, or departments, the line between them implies a sequence that does not exist. Use a bar chart or dot plot.
- Two time points. A two-point line is usually a slope chart. Treat it as a before/after comparison: direct labels, light axes, no unnecessary grid.
- Too many overlapping series. More than five lines on one panel often becomes a spaghetti chart. Use small multiples, filtering, or a highlighted focus series.
- Missing or irregular data. Connecting through gaps implies values were observed when they were not. Break the line, use a dotted segment, or annotate the gap.
- Part-to-whole composition. If the point is share of a total, an area chart, stacked area, or 100% stacked bar may be more appropriate.
Design Principles
The line chart is simple, but not forgiving. Most errors come from treating it as a decorative connector rather than a claim about continuity.
Use colour only when it has work to do
For a single measure, one colour is enough. Additional colours should identify meaningful series, forecast bands, interventions, or highlight states. A rainbow line chart usually asks the reader to decode decoration before they can read the data.
Bank the chart to reveal slope
William Cleveland's banking-to-45-degrees rule is a reminder that aspect ratio changes interpretation. Too wide and real movement looks flat. Too tall and ordinary noise looks dramatic. The goal is not always exactly 45 degrees, but the chart should give slope enough room to be read honestly.
Prefer direct labels to legends
Legends force visual lookup. The reader has to leave the line, decode colour, and return to the plot. For a small number of series, put the label at the right-hand end of each line. The label becomes part of the line.
Use point marks only when they earn their keep
Point marks can clarify sparse observations, exact event counts, and interactive targets. They are also visual noise. If the series is dense, the reader should see the trend first and use hover for exact values. If the observations are few, irregular, or individually important, marks help the reader see that the line is built from discrete measurements.
Do not force a zero baseline by habit
Bars encode value through length, so they usually need a zero baseline. Lines encode value through position and slope. A non-zero baseline is often appropriate when the analytical question is about change within a narrow range.
The caution is proportional interpretation. If a chart invites the reader to judge percentage change or magnitude relative to a whole, include zero or make the limited axis explicit.
Choose smoothing with care
Straight segments are faithful to observed points. Smoothed curves can look more natural, but they interpolate between observations and may imply a process that was never measured. Use monotone smoothing for genuinely continuous processes. Use straight segments for financial closes, survey waves, weekly reporting, or any series where the observations are the data.
Break the line for missing data
A missing value is not a zero and not a hidden observation. If the gap is material, break the line. If you interpolate, use a dotted segment and tell the reader. A continuous stroke is a claim that the path is known.
Annotate events, not every wiggle
The reader rarely needs every local maximum explained. They need the structural moments: a launch, policy change, outage, price move, campaign, or season boundary. Use a quiet reference rule and a short label.
Use area fill only when area has meaning
Filling under a line makes the chart heavier and encourages the reader to interpret volume. That can be useful for a single cumulative series. It is usually harmful for multiple overlapping series, where fills obscure each other and make the values harder to compare.
Anatomy
A line chart is a polyline plus axes. Everything else is scaffolding or annotation. Keep the scaffolding quiet enough that the trajectory remains the dominant object.
Related types
- Sparkline — a line chart reduced to its smallest useful form. No axes, no legend, usually embedded in text or tables.
- Area chart — a line chart with the region beneath filled. Useful when accumulated volume matters; risky for overlapping series.
- Slope chart — two time points connected by lines. Best for before/after comparison across categories.
- Bump chart — ranks over time. The y-axis is position, not magnitude.
- Step chart — values change in discrete jumps and remain constant between observations. Good for prices, inventory levels, policy states, and rates.
- Horizon chart — compresses dense time series vertically by folding bands of value. Useful for dashboards with many comparable series.
- Control chart — a line chart with statistical limits. Used for process monitoring and quality control.
Reading list
- Cleveland, W. (1993). Visualizing Data. Banking to 45 degrees and the perceptual basis of slope.
- Cleveland, W. & McGill, R. (1984). Graphical Perception. The foundational study on position along a common scale.
- Few, S. (2012). Show Me the Numbers. Time-series best practices and direct labelling.
- Tufte, E. (2006). Beautiful Evidence. Sparklines and data-rich time series.
- Heer, J. & Agrawala, M. (2006). Multi-Scale Banking to 45°. A formal treatment of aspect ratio for line charts.