You have the data, you have the tool, and now comes the million-dollar question: which chart type should I use? It seems like a simple decision, but it can completely change how your information is interpreted.
Choosing the right chart isn’t about aesthetics. It’s about clarity. A poorly chosen chart can confuse; a well-chosen chart can convince.
Before choosing: what do you want to show?
It all comes down to four main goals:
- Compare — Do you want to see how different categories stack up against each other?
- Show trends — Do you want to see how something changes over time?
- Show composition — Do you want to see the parts of a whole?
- Show distribution — Do you want to see how data is spread out?
Once you define your goal, the chart practically picks itself.
For comparison: horizontal or vertical bars
The bar chart is your best friend for comparisons. It’s familiar, easy to read, and works with almost any number of categories.
Use vertical bars when you have few categories (5-7) with short names. Use horizontal bars when you have many categories or long names — so the text doesn’t pile up.
Example: “organic traffic by country” with 15 countries → horizontal bars. “Conversions by quarter” with 4 quarters → vertical bars.
For trends: line charts
If your X-axis is time (days, weeks, months, years), a line chart is almost always your best bet. It shows evolution fluidly and lets you compare multiple series at once.
Golden rule: don’t put more than 4-5 lines in a single chart. After that, it becomes a spaghetti mess that nobody understands.
For composition: pie charts (with caution)
Yes, I know. Pie charts have a bad reputation. And for good reason — they’re constantly misused. But when you have 2 to 4 categories that add up to 100%, a pie chart can work well.
If you have more than 5 categories, use stacked bars or a treemap. The human eye is terrible at comparing angles, but excellent at comparing lengths.
For distribution: histograms or box plots
Want to see how your page load times are distributed? Or the prices of products in your e-commerce? You need a histogram or a box plot.
Histograms show frequency across ranges. Box plots show median, quartiles, and outliers at a glance. Both are powerful tools that many people don’t know about.
Quick cheat sheet
For quick reference:
- Compare categories → Bars
- Change over time → Lines
- Parts of a whole (few) → Pie or donut
- Parts of a whole (many) → Stacked bars or treemap
- Distribution → Histogram or box plot
- Relationship between variables → Scatter plot
- Geography → Choropleth map
If you want to explore more chart types with visual examples, I’m building Graficoteca: a catalog of 60+ visualization types in Spanish. Coming soon.
Remember: the best chart is the one your audience understands without needing an explanation. If you have to explain how to read it, it’s probably not the right chart.