Bar Chart
Horizontal bars excel at ranking. Long category names read naturally left-to-right, and dozens of items stack tidily down the page without sacrificing label clarity.
Charts That Actually Communicate
Sixteen chart types grouped by the question they answer — comparison, relationship, distribution, and composition. Pick the chart that fits the question, not the one that looks impressive.
Horizontal bars excel at ranking. Long category names read naturally left-to-right, and dozens of items stack tidily down the page without sacrificing label clarity.
The workhorse of comparison. Vertical columns let the eye judge length along a common baseline — one of the most accurate visual encodings for categorical data with short labels.
The clearest way to show change over a continuous variable — usually time. Slope tells you direction, steepness tells you rate, crossings tell you where one series overtakes another.
Plots several quantitative variables on spokes radiating from a shared center. Best for comparing the overall "shape" of a few items across the same set of metrics.
Each point is one observation across two variables. Reveals correlation, clusters, and outliers — the chart that lets the data's structure show itself, rather than imposing one.
A scatter plot with a third variable encoded as dot size. Reveals relationships across three dimensions at once — x position, y position, and magnitude.
A scatter plot whose points are joined in time order. Shows how the relationship between two variables evolves — tracing a path rather than a static cloud.
A grid where color intensity encodes a third value across two dimensions. Excellent for spotting density and pattern at a glance — calendar usage, correlation matrices, geographic clusters.
A bar chart for a single continuous variable, binned into ranges. Reveals the shape of a distribution — where the mass is, whether it's skewed, whether you're looking at one population or several.
A compact summary of a distribution: median, quartiles, range, and outliers — all in one mark. Indispensable when comparing the spread of many groups side-by-side.
A box plot crossed with a density curve — the width at each value shows where data clusters. Reveals multi-modal shapes a plain box plot would hide.
Every observation plotted along one axis, jittered to avoid overlap. Honest about sample size and outliers — you see the real data points, not a summary of them.
Parts of a whole, expressed as angle. Reliable only for two or three slices — beyond that, eyes can't distinguish small angular differences and a bar chart will read clearer.
A line chart with the area below filled in. Adds visual weight that emphasizes volume or cumulative magnitude — useful for showing how a total accumulates or how parts stack up.
Hierarchical proportions packed into a rectangle. Each block's area represents its value, and nesting shows containment — efficient for showing both share and structure in a single dense view.
Shows flow between states. Thickness of each ribbon represents quantity transferred — the best chart for tracing how a user, a dollar, or an electron moves through a system.
Charts are increasingly generated from a sentence — and the model can pick the encoding and narrate the insight for you.
Natural-language charting: "show revenue by region as bars" produces the chart. The prompt replaces the chart-type picker and the axis menus.
Given raw data, a model can suggest the right chart type and narrate the takeaway in plain language above it — a first draft a human still edits.