The Art & Design
of Data
This is the heart of the book. Choose the right chart for your question, pick colours everyone can read, strip away clutter, spot the tricks of misleading charts, and design infographics people remember, checked against an honest-chart checklist you can use for life.
- 00First Principlesp.2
- 01Choose a chartp.4
- 02Colourp.5
- 03Declutterp.6
- 04Honest or misleading?p.7
- 05The checklistp.8
- ✦Self-check · Make it · 5 Questions · Side Questp.9
You'll need: a notebook, grid paper, a pencil, a ruler, coloured pencils and, if you have one, a free spreadsheet or design app.
Principles
A first principle is a basic truth you can build on. The design of data rests on four: the question picks the chart, every mark should earn its place, colour is for meaning and beautiful must also be honest.
What do I already know?
What is necessary?
Everything in this module is built from four bricks. Rate yourself on each one, then check the example to see if your rating holds up.
What is the objective?
By the end, you’ll be able to choose a chart type to fit a question, pick sequential, diverging and colour-blind-safe palettes, declutter a chart using Tufte’s data-ink idea, spot truncated axes, 3-D pies, dual axes and cherry-picking, and check any chart or infographic against the honest-chart checklist.
- 01Choose a chart type.
- 02Pick a safe palette.
- 03Declutter a chart.
- 04Spot a truncated axis.
- 05Spot cherry-picking.
- 06Use the checklist.
The list on the left is ours. Now make it yours. What do you want out of this?
How much do you know about the art and design of data right now? Shade the boxes in pencil. You'll come back at the end (p.9) and shade it again in pen.
Why start this way? When you know what you already have (1), what you need (2) and where you're going (3), you learn faster and you notice the gaps. Every module on the Data Management Path starts with these same three questions.
Choosing a chart
Match the chart type to the question you’re asking.
The question decides
| Question | Chart |
|---|---|
| How do groups compare? | bar chart |
| How did it change over time? | line chart |
| What are the parts of a whole? | stacked bar, or a pie with few slices |
| Are two things related? | scatter plot |
| How spread out is it? | histogram, box plot |
| Where is it? | map |
Calgary vs Edmonton
Change over time, so a line chart. One solid and one dashed line help colour-blind readers.
Pencil before pixels
Designers sketch 3 or 4 rough versions on paper before opening any app. Ask: what’s the one thing a reader should notice in 5 seconds? Then pick the chart that makes that thing easiest to see, using what Lens 3 showed about position, length and angle.
- Which chart type do you see most in the news?
- What question would make you choose a map?
Match the chart
For each question, pick a chart and sketch it: How many days below zero did each month have? Is a warmer month also a wetter month? How did your height change since age 5?You want to show how Calgary’s temperature changed month by month. Which chart?
Colour and palettes
Use colour for meaning, and make sure everyone can see it.
Colour with a job
- Sequential: light to dark for low to high (rainfall)
- Diverging: two colours from a middle point (colder or warmer than normal)
- Categorical: clearly different colours for groups (cities)
The Okabe–Ito palette
In 2008, Masataka Okabe and Kei Ito published a set of 8 colours that people with common colour-vision deficiencies can tell apart. Many designers use it for categories.
Grey is a colour too
Use colour to make one thing stand out: one bright colour for the line that matters, grey for the rest. Too many colours turn a chart into a puzzle. Keep the same colour for the same thing on every chart (Calgary is always green), and check it in greyscale (Lens 3).
- Which colours would you use for hot and cold?
- How many colours is too many?
Recolour it
Draw a bar chart of 4 made-up favourite seasons with 4 random bright colours. Redraw it with grey bars and one highlight colour for the winner. Which version tells the story faster?Which palette fits “colder or warmer than normal”?
Decluttering
Remove anything that doesn’t help the reader see the data.
Every mark earns its place
In 1983, Edward Tufte’s book The Visual Display of Quantitative Information argued that most of the ink in a chart should show data. He called needless decoration chartjunk: heavy gridlines, 3-D effects, shadows, clip art and busy backgrounds.
Before and after
| Before | After |
|---|---|
| 3-D bars | flat bars |
| legend box | labels on the lines |
| dark gridlines | light, or none |
| title “Data” | title with the finding |
| rainbow colours | grey plus 1 highlight |
| no source | source and licence |
Delight with a purpose
Decluttering doesn’t mean boring. Great infographics use striking type, a clear layout, illustration and white space, and Nightingale’s rose (Lens 5) is beautiful. Some studies suggest a memorable picture can help people recall a chart. The rule: decoration must never hide or bend the data.
- Can decoration ever help a chart?
- Where is the line between beautiful and distracting?
Find the junk
Find a cluttered chart (an ad, a report or a slide). List every element that doesn’t show data. Redraw it on grid paper with only what helps the reader, and add a finding title.What did Tufte call needless decoration in charts?
Honest vs misleading charts
Bars start at zero, and every trick changes what readers believe.
Same data, two stories
Total 2024 precipitation (mm). On the right, Calgary’s bar looks about 8 times taller; really it’s about 1.5 times as much.
Watch for these
- Bars not starting at 0: bar length must show the value
- 3-D pies: front slices look bigger
- Two y-axes: you can make any two lines cross
- Cherry-picking: showing only Calgary’s dry October
- Area for 1-D data: doubling a picture’s width and height makes it look 4 times bigger
When not starting at zero is fine
A bar shows a value by its length, so it must start at zero. A line shows change, so its axis can zoom in, as long as it’s clearly labelled; temperature in °C has no natural zero anyway. The test: would an honest reader get the right idea of the size of the difference?
- Have you seen a chart that made a small difference look huge?
- Is a misleading chart a kind of lie?
Make it lie, then fix it
Use the dataset at the end of this book to draw one misleading chart on purpose (truncate an axis or cherry-pick months). Then draw an honest version. Show both to someone: what did each make them believe?Why must a bar chart’s axis start at zero?
The honest-chart checklist
Beautiful charts must pass every honest check.
Use it on every chart
A one-page story
- One question, answered in the headline
- One big number readers will remember
- 1 to 3 honest charts, not 10
- A clear reading path: top to bottom
- Source line at the bottom
Design moves
- Pick 2 fonts at most: one bold, one plain
- Use white space to group things
- Repeat colours: Calgary always green
- Use simple icons that match the data, not clip art
- Draw it by hand first
- Which check would people forget most often?
- Should news sites have to pass a checklist like this?
Score 3 charts
Find 3 charts (news, ads, textbooks). Score each one out of 12 with the checklist. Which check failed most often? Fix the worst chart.A chart passes 11 checks but has no source. Is it ready to share?
Quick self-check
Six questions, covering every stop. Show your thinking. Answers are upside down at the bottom; no peeking till you're done.
1 · Which chart fits change over time?
2 · What is a diverging palette for?
3 · What did Tufte call needless decoration?
4 · A bar chart’s axis starts at 350. What’s wrong?
5 · Name one other misleading-chart trick.
6 · Name 3 checks from the honest-chart checklist.
Re-shade your knowledge meter
How much do you know about the art and design of data now? Flip back to p.3 and shade the meter again in pen (or shade this one). Did any of the four bricks on p.2 move from "Not yet" to "Got it"?
Do a chart makeover
Can you make a chart honest and beautiful?
1 · Plan
Find a confusing or misleading chart (an ad, a news story, a report, or the truncated chart in section 04). Run it through the honest-chart checklist, then redesign it so it is honest, clear and beautiful.
- Chart found
- Source noted
- Checklist run
- Sketches made
2 · Checks it failed
| Check | Problem | My fix |
|---|---|---|
3 · Must-haves
- Finding title
- Honest axis
- Safe colours
- Source line
4 · What I found
Show the before and after to 2 people. What did each version make them believe? Which change made the biggest difference? Is your version more beautiful, too?
Critique the chart, not the person who made it. If you share a makeover of someone else’s chart, credit the original.
5 Questions
I Still Have
This module didn't answer everything. Good. Write 5 new questions it didn't answer, one of each type below. There are no wrong questions, only ones nobody's asked yet.
★ Star the one question you'd most like answered. You'll need it on the next page.
1What I'll find out
Break your big question into 2 or 3 smaller ones you could actually answer.
2Where I'll look
3How I'll share it
Who I'll share it with:
Done by: