Leo+DadMade for Leo
Picturing Data
Rung 1 of 4 · Discover

Where the Right Chart Comes From

Before any maths, let's see the one move that turns a wall of numbers into something you can read — because the same data drawn the wrong way tells a lie.

NESA SC4-DA1-01 A picture beats a list

Play The same little dataset, drawn two ways — one sensible, one silly. Pick the chart that tells the truth.
🎧
Audio WalkthroughDad & Leo, Two Minutes — Coming Soon
Video ExplainerComing Soon

Here's the whole idea in one breath: a picture beats a column of numbers — but only if the picture matches the kind of data you've got. Numbers in a list hide their pattern; the right chart hands it to you in a second. The wrong chart hides it again, or makes up a pattern that was never there.

Stare at a Column of Numbers

Imagine a table: the daily maximum temperature for two weeks — fourteen numbers in a row. You can read it, but you can't really see it. Is it warming up? Cooling? Was Tuesday the odd one out? Your eye has to crawl down the list and hold every number at once, and it can't. Draw those same fourteen numbers as a line that rises and dips across the page, and the answer jumps out before you've even thought about it. Nothing changed in the data. Everything changed in how fast you could read it.

Say it plainly: a chart doesn't add anything to your data — it just lets your eyes do the work your brain was struggling with. The skill isn't drawing the chart; it's picking the right chart for the kind of data you have.

Match the Data to the Chart

There are four everyday data shapes, and each one has a chart that fits it like a key in a lock. Categories — things sorted into named groups — want a bar or column chart, one bar per group, taller means more. Parts of a whole — slices that add to one complete thing — want a pie chart, but only ever when the parts genuinely make up 100% of something. Change over time — the same thing measured again and again as the clock ticks — wants a line chart, time along the bottom and the trend running up, down or steady. And a spread of measurements — lots of readings of one quantity — wants a dot plot or histogram to show where the values bunch and how far they scatter.

Drag the chooser in the toy and you'll feel it: put change over time into a pie chart and it turns to nonsense — a pie can't show an order, so the days lose their sequence and the trend vanishes. Match it to a line and the story comes straight back. That mismatch is the single most common data mistake, and you can now spot it in a heartbeat.

Us, Thinking Out Loud

You measured the temperature every day for a fortnight. Which chart — and why not a pie?

When would a pie chart actually be the right call?