Every chart type makes an argument. A line chart argues that the data is continuous and that the trend matters. A bar chart argues that comparisons between discrete categories matter. A scatter plot argues that the relationship between two variables matters. When you choose a chart type, you are not making a stylistic decision—you are making an editorial one.
This is why choosing the wrong chart is not merely an aesthetic failure. It is a communication failure. A pie chart showing time-series data does not just look wrong—it actively prevents the viewer from seeing the trend. A line chart comparing five unrelated categories does not just confuse—it fabricates a relationship that does not exist.
The Four Questions
Four questions that determine the right chart type for any dataset.
Before selecting a chart, ask four questions about your data: What is the relationship you want to show? How many variables are involved? Is the data categorical or continuous? What comparison matters most to your audience?
The answers map to chart families:
- Comparison across categories → bar chart, grouped bar, dot plot
- Change over time → line chart, area chart, slope chart
- Part-to-whole relationship → stacked bar, treemap, waffle chart
- Correlation between variables → scatter plot, bubble chart, heatmap
- Distribution of values → histogram, box plot, violin plot
Encoding Channels
Humans decode position most accurately, followed by length, angle, area, and finally color.
Not all visual encodings are created equal. Research in perceptual psychology has established a clear hierarchy: position along a common scale is the most accurately perceived, followed by length, angle, area, and finally color saturation. This is why bar charts (position + length) are almost always more readable than pie charts (angle + area).
Understanding encoding channels transforms chart selection from guesswork into science. When you need precision—when the viewer must be able to compare values accurately—use position. When you need to show general patterns or categories, color is appropriate. The encoding should match the precision required by the task.
“The best chart is the one that makes the right answer obvious and the wrong answer impossible.
Common Mistakes
The most pervasive chart crime is dual-axis charts. By placing two different scales on the same plot, you allow any two trends to appear correlated—simply by adjusting the axis ranges. Dual-axis charts are the data visualization equivalent of lying with statistics, and they should be avoided in almost all cases.
Other common mistakes include truncated y-axes (which exaggerate differences), 3D effects (which distort area perception), and rainbow color palettes (which create false boundaries in continuous data). Each of these transforms a neutral data display into an opinionated—and often misleading—argument.
Choosing the right chart is not about memorizing rules. It is about developing an intuition for how visual encodings map to human perception. When you internalize this mapping, chart selection becomes effortless—and your data visualizations become instruments of genuine understanding rather than decoration.