PowerPoint Chart Type Selection Guide
Choosing the wrong chart type is the most common data visualization mistake in presentations. A pie chart with 12 slices, a line chart for categorical data, a 3D bar chart for anything — each makes the data harder to understand, not easier.
This guide maps data relationships to chart types so you always pick the right one.
The Decision Framework
Start with your data's story, not its shape. Ask: what relationship am I showing?
| Relationship | Best Chart Types |
|---|---|
| Comparison across categories | Column, bar, grouped bar |
| Change over time | Line, area, column |
| Part-to-whole | Pie (≤5 slices), stacked bar, treemap |
| Distribution | Histogram, box-and-whisker, scatter |
| Correlation | Scatter plot, bubble chart |
| Ranking | Horizontal bar (sorted) |
| Flow or accumulation | Waterfall, funnel |
| Geographic patterns | Map/geo chart |
| KPI vs. target | Bullet chart |
| Composition of composition | Treemap, sunburst |
Comparison Charts
Column Charts
Best for comparing values across a small number of categories (3–12). Vertical orientation is natural for showing magnitude.
Bar Charts
Horizontal bars work better when category labels are long (country names, product names) or when you have many categories (12+). Also better for ranking data — sort bars by length.
Grouped and Stacked Bars
Grouped bars compare subcategories within categories. Stacked bars show part-to-whole within each category. Don't use stacked bars if you need precise comparison of the stacked segments — only the bottom segment shares a common baseline.
Trend Charts
Line Charts
The default for time-series data. Use when you have continuous data over time and want to show direction, rate of change, and patterns. Limit to 4–5 lines before the chart becomes unreadable.
Area Charts
Like line charts, but the filled area emphasizes volume or magnitude. Stacked area charts show how parts contribute to a total over time. Use with caution — overlapping areas can obscure data.
Proportion Charts
Pie and Doughnut Charts
Use only when showing parts of a whole with 5 or fewer slices. Humans are poor at comparing angles — if precision matters, use a bar chart instead. Doughnut charts free up the center for a total or label.
Treemaps
Show hierarchical part-to-whole relationships using nested rectangles. Better than pie charts for many categories. Each rectangle's area represents its value.
Distribution and Correlation
Scatter Plots
Show the relationship between two variables. Each point represents one observation. Add a trend line to highlight correlation direction and strength.
Bubble Charts
Scatter plots with a third variable encoded as bubble size. Use sparingly — the human eye is poor at comparing circle areas precisely.
Box-and-Whisker
Show the distribution of a dataset: median, quartiles, and outliers. Ideal for comparing distributions across groups.
Specialized Charts
Waterfall Charts
Show how a starting value increases and decreases through a series of intermediate values to arrive at a final total. Perfect for financial bridges (revenue to profit).
Funnel Charts
Show progressive reduction through stages (leads → opportunities → deals). The narrowing shape reinforces the dropping numbers.
Radar/Spider Charts
Compare multiple variables for one or a few items. Works for competency assessments, product comparisons across dimensions. Avoid with more than 8 axes.
Bullet Charts
Show a single KPI against a target, with qualitative ranges (poor/good/excellent). The most space-efficient KPI visualization.
Pareto Charts
Combine a bar chart (sorted descending) with a cumulative line to identify the vital few factors that account for most of the effect.
The Chart Type Cheat Sheet
| Your data looks like... | Use this |
|---|---|
| Sales by region | Horizontal bar (sorted) |
| Revenue by quarter | Column or line |
| Market share | Pie (≤5) or horizontal bar |
| Revenue bridge | Waterfall |
| Sales pipeline stages | Funnel |
| Satisfaction survey results | Horizontal bar or radar |
| Two variables, looking for correlation | Scatter |
| KPI vs. target | Bullet |
| Failure cause analysis | Pareto |
Universal Rules
- Never use 3D charts. They distort perception and add zero information.
- Sort bar charts by value unless there's a natural order (time, geography).
- Start Y-axes at zero for bar and column charts. Truncated axes exaggerate differences.
- Label directly on the chart instead of using a separate legend when possible.
- One chart per slide for important insights. Two maximum.