04 SEEING · How eyes read a picture
Dense, not decorated.
You will learn how to show a lot of data clearly, with structure instead of decoration. It takes 3 minutes.
Watch the counter
The clip starts with a chart that shows four numbers. Look at the counter in the corner as it goes. It counts the numbers on screen.
The name for it
THE NAME FOR IT
Edward Tufte's information design
Edward Tufte · The Visual Display of Quantitative Information, 1983
Show the data, invite comparison, label directly, mix words and graphics, and spend ink on information, not decoration.
Four numbers need no chart
The clip's first chart: Q1 42, Q2 51, Q3 47, Q4 58. That sentence holds all of it. The 3D bars, gradients, heavy grid, clip art and legend added ink and no information.
But stripping decoration is only half of Tufte. The other half: once the ink is honest, show more.
288 numbers, one look
WEEKLY VISITORS PER STORE, LAST 24 MONTHS · EXAMPLE DATA
What you can see in a few seconds:
- Most stores peak in the same season. The shared scale makes the timing comparable.
- Three months are missing at one store. The note sits on the dip, not in a footnote.
- Big stores and small stores are obvious. The scale is the same, so height means the same thing everywhere.
- No legend. Names sit on the panels. The colors have one job each: red is a low, green is a high.
The habits
- Show the data. Spend ink on numbers, not on frames, shadows and backgrounds.
- Invite comparison. Put related things side by side, on the same scale.
- Label directly. Put names on the lines and bars. A legend makes the eye travel back and forth.
- Mix words and graphics. Put the explanation where it applies.
- Add data, not decoration. If the chart has room, give it more context, not more effects.
"Clutter and confusion are failures of design, not attributes of information."
Edward Tufte, Envisioning Information (1990)
TAKEAWAY
Not less information.
More, with structure.
Ask your agent
Make this chart in Tufte's style. Remove: 3D, gradients, shadows, background fills, heavy gridlines, legends, decorative icons. Label lines and bars directly. Put any explanation as a short note next to the data it explains. If there are many series, use small multiples on one shared scale instead of one tangled chart. Then ask: what context could this chart show in the space it now has? Add it if it helps the reader compare. Data and the question the reader has: [paste here]
What good output looks like: a plain chart with more data in it than before, direct labels and no legend.
Sources
- Edward R. Tufte, The Visual Display of Quantitative Information, 1983. The source of data-ink, chartjunk and small multiples.
- Edward R. Tufte, Envisioning Information, 1990. The quote above.
- The store data is invented example data.