Bar Chart vs Line Chart: Match the Graphic to the Question
TLDR
Use a bar chart to compare discrete categories, such as products, regions or survey responses. Use a line chart to show how a value changes across a continuous, ordered scale—most often time. Bars normally need a zero baseline because their lengths encode values. Lines can use a non-zero baseline when the narrower range is clearly shown and genuinely helps readers see meaningful movement.
The deciding question is not whether the dataset contains dates. Ask whether readers need to compare separate values or follow the direction, pace and shape of change. Monthly sales could use either format: bars emphasize the amount in each month, while a line emphasizes the trend across months.
The basic choice: comparison or progression
| Question the chart should answer | Usually choose | Reason |
|---|---|---|
| Which category is larger? | Bar chart | Bar lengths make separate values easy to compare. |
| How has a measure changed over time? | Line chart | Connected positions emphasize direction and the shape of change. |
| What was the amount in each period? | Bar chart | Each period is presented as a distinct amount. |
| How do several time series move together? | Line chart | Lines usually make relative movement clearer than grouped bars. |
| How much did each category increase or decrease? | Bar chart around zero | Bars can extend in opposite directions from a common reference line. |
Bar charts work well for nominal categories that are separate rather than continuous. Each bar represents one category, and a useful order—such as highest to lowest or a natural sequence—can make the comparison faster to read. This is also why bars are generally more effective than slices when readers must judge similar values, as explained in our comparison of pie charts and bar charts.
Line charts are designed for an ordered progression. Time is the most familiar x-axis, but another ordered numeric variable, such as experimental dose, can also work. Connecting observations directs attention to direction, pace and overall shape. Fundamentals of Data Visualization describes this ordered relationship as the defining use of a time-series line chart.
Do not connect categories that have no continuous relationship
Placing categories from left to right does not turn them into a continuous series. A line connecting brands, industries or departments implies that adjacent categories have a meaningful order and that the intervals between them matter. Those assumptions usually do not hold.
Suppose a chart lists retail, manufacturing and healthcare. A rising line across those labels visually suggests movement from one sector to the next, even though there is no measured path between them. Separate bars communicate the intended question more honestly: which sector has the higher value?
Ordered categories are a closer call. Age bands, satisfaction levels and income brackets have a sequence, but they remain grouped ranges rather than exact points on a continuous scale. Bars are often the safer choice when the main task is comparing those groups.
Time data can use bars or lines
Dates do not automatically require a line chart. Bars can represent equal-duration periods such as months, quarters or years, especially when the amount in each period matters more than the path between them. The UK Government Analysis Function’s chart guidance notes that bars can show time-based values, while lines generally make multiple time series easier to compare than grouped bars.
Consider quarterly revenue. Use columns if the question is “How much revenue was recorded in each quarter?” Use a line if the question is “Is revenue rising, falling or changing direction?” The numbers may be identical; the graphic changes the analytical emphasis.
For composition over time, neither an ordinary bar nor a single line may be sufficient. A stacked display can show how a total is divided, but its moving baselines make middle categories difficult to compare. Our guide to stacked area chart limitations explains that tradeoff.
Preserve actual spacing for irregular intervals
A line chart can show observations collected at irregular dates, but the x-axis should preserve elapsed time. Measurements taken on January 1, January 2 and January 20 should not appear equally spaced. Equal visual gaps would turn one day and 18 days into the same apparent interval.
The Office for National Statistics line-chart guidance recommends displaying markers when they help reveal individual observations and their frequency. Markers are particularly useful for irregular series because they show where measurements actually occurred.
A segment between two points is a visual connection, not proof that every intermediate value was measured. If an expected observation is missing, leave a gap rather than drawing across it. The same principle applies to a methodological break or other discontinuity in the series.
Bar charts should normally start at zero
On a linear scale, a bar chart’s value axis should begin at zero. Bars encode magnitude through the length and area of a filled shape measured from a baseline. Cropping that baseline changes the apparent length ratio and can make a small difference look large.
For example, values of 98 and 100 differ by about 2%. If bars begin at 97, one visible bar would be three times as long as the other. The labels might remain numerically correct, but the geometry would exaggerate the practical difference.
When charting changes rather than levels, zero can sit inside the plotting area. Negative bars extend to one side and positive bars to the other, making the direction and magnitude of change visible from the same reference point.
Line charts can use a non-zero baseline—with care
A line chart encodes each value primarily by its vertical position, not by the length of a filled mark from zero. It therefore does not always need a zero baseline. The Office for National Statistics permits cropped line-chart axes when a narrower range makes meaningful movement legible, while noting that zero is preferable when it does not obscure the trend. ONS guidance on axes and gridlines also recommends clearly separating the axis start from the first plotted value when zero is omitted.
A cropped axis changes the emphasis. It can reveal small but important variation, yet it can also make modest movement appear dramatic. Broken axes add another layer of interpretation and may make differences harder to judge. Use a clearly labeled scale, retain enough context to show the magnitude of the values, and avoid framing routine noise as a major swing.
Check what the chart hides
Choosing the correct chart family does not fix weak data. Missing observations, revisions, methodological breaks and sampling uncertainty can undermine an apparently clean comparison or trend. When uncertainty could change the conclusion, show an interval or another appropriate representation rather than presenting every estimate as exact.
Clutter also changes the choice. The Government Analysis Function suggests aiming for no more than four lines as a guideline, not an absolute rule. Grouped bars likewise become difficult to scan as categories and series multiply. If readers must decode a crowded legend and trace many overlapping marks, splitting the display into small multiples may be clearer.
The practical rule
Choose the chart from the relationship you want readers to see. Use bars for separate categories and period-by-period amounts. Use lines for movement across a genuinely ordered, continuous scale. Preserve real time spacing, show gaps where observations are missing, and treat the zero baseline differently because bars encode length while lines encode position.
If both formats seem defensible, write the chart’s question as a sentence. “Which value is larger?” points to bars. “How is the value changing?” points to a line. That single distinction resolves most bar-chart-versus-line-chart decisions before styling begins.
References
- New Bar Chart Reference Page – The Data Visualisation Catalogue Blog
- Fundamentals of Data Visualization
- How to Spot Misleading Charts, a Checklist
- Data visualisation: charts – Government Analysis Function
- Data over time: Line chart – Data visualisation – Service manual – Office for National Statistics
- Chart details: Axes and gridlines – Data visualisation – Service manual – Office for National Statistics
- Fundamentals of Data Visualization
- Overall considerations: Showing uncertainty in charts – Data visualisation – Service manual – Office for National Statistics
