Why Charts Exist

The previous chapter defined what a visualization is. This chapter asks a related but distinct question: why do we need charts at all? A table of numbers contains all the same information. Why do humans, in every culture and every era, keep inventing new ways to draw the data?

The answer is the central reason data visualization is a discipline.

Tables make you read; charts let you see

Take 30 seconds to look at this table. It shows life expectancy in seven countries in two years (1952 and 2007).

Now answer:

• Which country had the largest gain?
• Which countries were already over 60 in 1952?
• Is there a country that grew faster than China?

You can answer all three from the table — but it takes mental effort. Your eye has to find each number, your brain has to remember what you found, and you have to do subtractions in your head.

Now look at the same data as a chart:

Now the questions are trivial:

• Largest gain = steepest line. China and Brazil leap up; Afghanistan is the most modest gain.
• Already over 60 in 1952 = the dots above 60 on the left. Germany, Japan, Norway.
• Grew faster than China = a line steeper than China's. Brazil is competitive; none is clearly steeper.

You did not consciously compute any of those. Your eye did the work.

What chart is really doing

Several things are happening at once that the table cannot do:

1. Spatial proximity lets your eye compare adjacent values.
2. Lines show direction-of-change at a glance.
3. Color binds the same country together across the chart.
4. Aligned axes let your brain compare all countries to all other countries simultaneously.
5. Pattern recognition — your visual system spots the upward trend in essentially zero conscious time.

These are all things the visual cortex is built for and the conscious arithmetic part of the brain is terrible at.

When tables are actually better

Charts are not always the right answer. Tables win when:

• The reader needs to read exact values (financial statements, scientific reports).
• The data has many columns of different kinds (a table of product specs).
• The reader will look up specific rows ("what's the price of SKU 7?").
• The data is sparse: just a few values that don't share comparable units.

A good rule of thumb:

• Looking up? Use a table.
• Comparing or finding patterns? Use a chart.

Many real reports have both — a chart for the story and a table below it for the receipts.

Charts as cognitive offloading

Researchers in cognition use the phrase "cognitive offloading" for tools that move some of the work of thinking out of the brain and into the world. A grocery list is cognitive offloading: you don't try to remember every item, you read them off paper.

A chart is the same idea, applied to data. The chart "remembers" the data values for you, and arranges them so that your visual system can do certain comparisons automatically. The brain only has to do the interesting part of the thinking.

This is also why charts are so good for collaboration. If you and a colleague are both looking at the same chart, you are sharing the same offloaded thinking. You can point at a peak and say "look at that." You can't do that with two columns of numbers in a spreadsheet.

Why charts are sometimes worse than tables

Beginners sometimes hear "charts are better than tables" as a universal rule. They aren't. Charts have failure modes too:

• They round. A bar chart with values 99 and 101 looks identical to one with 100 and 100. The eye can't see a 2% difference; a table can.
• They mislead with bad encodings. A 3-D pie chart can make the same number look bigger or smaller depending on the angle.
• They invite over-interpretation. Seeing a "trend" in three data points is easy — and often wrong.
• They impose a frame. The author's choice of axes, colors, and scales shapes how the data is read.

The honest rule isn't "charts beat tables." It's "charts and tables are different tools for different jobs. Pick the one that answers the question."

Check your understanding

QuestionSelect one

Why are charts often easier to read than tables of numbers?

Charts always show more data than tables.

Charts use less memory than tables.

Charts let the human visual system perceive patterns in parallel — comparisons of position, length, and slope happen almost instantly without conscious effort.

Charts are required by law.

QuestionSelect one

Which of the following is a case where a table is probably better than a chart?

Comparing trends in revenue across 12 quarters.

Showing the distribution of customer ages.

An invoice listing each line item with its exact price.

Showing how two variables relate to each other.

QuestionSelect one

Which is NOT a legitimate failure mode of charts that tables don't share?

Rounding away small but important differences.

Inviting over-interpretation of short trends.

Imposing the author's framing through axes and colors.

Always being slower to read than tables.

QuestionSelect one

What does "cognitive offloading" mean in the context of data visualization?

Using a computer to do calculations.

Writing notes by hand.

Moving some of the work of thinking — remembering values, comparing them — out of the brain and into an external representation like a chart.

Forgetting things on purpose.