Learn data visualization basics

Discover how to turn raw numbers into honest, easy-to-read charts.

Time:: 60 min
Cost:: free
Difficulty:: moderate
Where:: Indoors
Best:: Solo

Data visualization basics are less about software than about a single decision: what question is this chart answering? Comparison, composition, distribution, and change over time each have a natural chart type, and picking the wrong one is the most common reason a chart fails to communicate anything at all.

Why learn data visualization basics is worth an hour

Charts are read pre-attentively, meaning the eye extracts the pattern before conscious thought engages. That is why encoding matters so much: length and position are read accurately, area and color intensity are not. Matching your question to an encoding people read well does most of the work, before any styling decisions arise.

How to learn data visualization basics step by step

1
Write the question the chart answers
One sentence. If you cannot state it, no chart type will rescue the visualisation.
2
Match the question to a form
Comparison uses bars, change over time uses lines, distribution uses histograms, relationship uses scatter plots. Composition is usually a bar too.
3
Remove everything decorative
Gridlines, borders, drop shadows, background fills. Every non-data element competes with the data for attention.
4
Start numeric axes at zero for bars
Truncating a bar axis exaggerates differences dramatically. Line charts showing change may legitimately do otherwise.
5
Label directly where you can
Put labels on the lines rather than in a legend. Every legend lookup costs the reader attention.

Tips to get more out of it

Avoid pie charts with more than three slices. People compare angles poorly, and a bar chart answers the same question better.
Use color to encode meaning, never for decoration. Check that the chart still works in greyscale.
Give the chart a title stating the finding, not the variables. "Sales fell in Q3" beats "Sales by quarter".

Ways to mix it up

If you present regularly, spend the hour building a small reusable template with your fonts and colors already set, which removes the styling decision permanently. If you consume more charts than you make, spend it learning to spot misleading axes and cherry-picked ranges instead.