Bar or line? Pie or donut? Horizontal or vertical? Stacked or grouped? The number of chart options is overwhelming, and picking the wrong one can completely misrepresent your data. Here's a decision framework that actually helps.
The Chart Decision Tree
Start with one question: what relationship are you showing?
- Comparison → Bar chart (vertical for few items, horizontal for many)
- Change over time → Line chart (continuous) or bar chart (discrete periods)
- Proportion → Stacked bar, treemap, or pie (only if 2-4 segments)
- Correlation → Scatter plot
- Distribution → Histogram, box plot, or violin plot
- Geographic → Map (choropleth or bubble)
The AI Chart Generator automates this decision. Paste your data, describe what you want to show, and it selects the appropriate chart type and generates it.
The Pie Chart Debate
Pie charts are the most misused chart type in existence. According to Tableau's research, humans are terrible at comparing angles and areas — which is exactly what pie charts ask us to do.
When pie charts work: 2-3 segments where one is clearly dominant (like "75% yes, 25% no"). When they don't: anything with more than 4 segments, or when segments are similar in size.
Alternative: a horizontal bar chart sorted by value. Always clearer, always more precise.
Formatting Rules That Make Charts Readable
- Title: State the insight, not the data. "Sales grew 23% in Q3" not "Q3 Sales Data"
- Axis labels: Always include units. "$K" or "thousands" or "%"
- Gridlines: Light gray, minimal. Remove if the chart is simple enough without them.
- Legend: Place it close to the data it describes. Better yet, label the data directly.
- Source: Always cite your data source. Builds credibility.
Common Mistakes
Truncated Y-axis: Starting the Y-axis at 50 instead of 0 makes a 5% change look like a 50% change. Always start at 0 for bar charts. Line charts can start higher if you note it clearly.
Dual Y-axes: Almost always misleading. Two different scales on the same chart lets you imply correlations that don't exist. Use two separate charts instead.
Too many colors: If your legend has more than 6 items, your chart has too many categories. Group the small ones into "Other."
Data Preparation
Good charts start with clean data. Use the Data Cleaning Tool to remove duplicates and fix formatting. The Data Visualizer can suggest the best chart type. For complex analysis, the Report Generator combines charts with narrative context.
As HBR's visualization guide puts it: "The goal of a chart is not to show data. It's to show an insight."
Generate the right chart for your data.
Try the Chart Generator →