Data visualization
How to choose the right chart and tell the right story
Most data visualizations fail not because of the data, but because of the chart choice. Learn which chart type to use when, how to eliminate chartjunk, and how to tell a clear data story.
Why visualization matters more than the data
The same data can tell completely different stories depending on how it is visualized. A bad chart hides insight; a great chart makes it obvious.
The goal of visualization is not to show data — it is to change what someone understands. Every design decision in a chart — the axis range, the color, the title, the chart type itself — either helps or hinders that goal. Nothing is neutral.
Chart type by purpose
Pick the chart type based on what you are trying to show, not what looks impressive. The wrong chart type forces the audience to do work the chart should do for them.
The chart mistakes that ruin credibility
These mistakes are noticed immediately by anyone with analytical experience and undermine trust in the entire analysis, regardless of how good the underlying work is.
Truncated Y-axis
Starting the Y-axis at something other than zero makes small changes look dramatic. Show the full axis unless there is a strong reason not to.
Pie charts with too many slices
More than 4–5 slices and it becomes unreadable. Use a bar chart instead.
Dual axis charts
Two Y-axes create the illusion of correlation. Use separate charts unless the relationship is the story.
3D charts
Add visual noise without adding information. Avoid always.
Chart without a title that states the insight
A title that says 'Monthly Revenue' is a label. A title that says 'Revenue Growth Accelerated in Q3' tells the story.
The data story formula
This structure — from Barbara Minto's Pyramid Principle — works for dashboards, reports, and presentations. It matches how humans process information: situation first, then problem, then answer.
Context → Complication → Question → Answer
Context
What is the situation? (Where we are)
Ground the audience before showing any numbers. What is the business situation? What question is being answered? Without this, the data floats in a vacuum.
Complication
What changed or is not working? (The problem)
This is the finding. Name what is unexpected or concerning. If nothing is surprising, question whether the analysis is worth presenting.
Question
What needs to be answered?
State the specific question the data will answer. This focuses the audience and signals that a recommendation is coming.
Answer
The data insight and recommendation.
Give the finding directly — the number, the trend, the comparison — followed by a specific recommended action. Uncertainty is not a reason to omit a recommendation.
Color in data visualization
Color is the most misused element in data visualization. Most charts use color for decoration. Color should encode information — every color difference should mean something.
Rule: use color to encode information, not for decoration. If two things are the same color in your chart, the audience assumes they are the same kind of thing.
Keep building
Learn data storytelling
Choosing the right chart is the first step. Turning that chart into a compelling narrative that drives decisions is the skill that separates analysts from data storytellers.
Learn data storytelling