- Home
- Chart guide
Chart guide
One page per chart type. Each starts with what the chart is good at, then covers the cases where something else would do the job better, and ends with the mistakes we see most often.
Bar chart
When a bar chart is the right choice, how to order and label the bars, why the axis has to start at zero, and when grouped bars stop working.
Line chart
When to use a line chart, how to highlight one series among many, why dual axes mislead, and how to handle missing data and spaghetti charts.
Scatter plot
How to read and build a scatter plot: spotting correlation, handling outliers and overplotting, adding trend lines, and using bubble size as a third variable.
Histogram
What a histogram shows, how to choose bin width, how it differs from a bar chart, how to read skew, and how to compare two distributions fairly.
Heatmap
How to build a readable heatmap: when a grid of colored cells works, sequential vs. diverging color scales, ordering rows and columns, and legends.
Stacked bar chart
When stacked bars work, why only the bottom segment is easy to compare, when to use 100% stacked bars, and when grouped bars or small multiples are better.
Pie and donut charts
When a pie chart is a fine choice, why charts with many slices fail, how donut charts differ, and which alternatives to use for part-to-whole data.
Treemap
When a treemap works for nested part-to-whole data, why area is hard to compare, how to label small rectangles, and when sorted bars are the better choice.