Graphical design: Scientific neutrality and Tufte

Lecture 14

Dr. Benjamin Soltoff

Cornell University
INFO 3312/5312 - Spring 2026

March 13, 2026

Announcements

Announcements

  • Project 01
  • Project 02

Edward Tufte

Image credit: Edward Tufte

Edward Tufte

  • Pioneer in the field of data visualization
  • PhD in political science
  • Professor emeritus of political science, statistics, and computer science at Yale
  • Collaborator with John Tukey
  • Self published 5 books on information design and data visualization

Principles of graphical excellence

Above all else show the data.

  • Show the data
  • Induce the viewer to think about the substance rather than about methodology, graphic design, the technology of graphic production, or something else
  • Avoid distorting what the data have to say
  • Present many numbers in a small space
  • Make large data sets coherent
  • Encourage the eye to compare different pieces of data
  • Reveal the data at several levels of detail, from a broad overview to the fine structure
  • Serve a reasonably clear purpose: description, exploration, tabulation, or decoration
  • Be closely integrated with the statistical and verbal descriptions of a data set

Source: The Visual Display of Quantitative Information, ch 1

Data-Ink

Goal is to maximize the data-ink ratio

\[\text{Data-ink ratio} = \frac{\text{data-ink}}{\text{total ink used to print the graphic}}\]

Proportion of a graphic’s ink devoted to the non-redundant display of data-information

What parts of a graph are data-ink?

Identifying data-ink

Ink on the chart

  • Data points
  • Axis ticks
  • Axis tick labels
  • Axis labels
  • Background
  • Grid lines

Reduce non-data ink

#| include: false

library(tidyverse)

library(ggthemes)

penguins <- drop_na(penguins) |>
  mutate(species = fct_infreq(species))

p <- ggplot(data = penguins, mapping = aes(x = flipper_len, y = body_mass)) +
  geom_point() +
  labs(
    x = "Flipper length (mm)",
    y = "Body mass (g)"
  )
#| warning: false
#| min-lines: 6
#| max-lines: 8
p

What about a bar graph?

Reduce non-data ink

#| warning: false
#| min-lines: 6
#| max-lines: 8
ggplot(data = penguins, mapping = aes(x = species)) +
  geom_bar() +
  labs(
    x = "Species",
    y = NULL,
    title = "Number of penguins by species"
  )

Application exercise

ae-13

Instructions

Redesign a boxplot to maximize the share of data-ink and reduce unnecessary duplication

Maximized data-ink

#| include: false
wdi <- read_csv("data/wdi.csv") |>
  filter(year == max(year)) |>
  drop_na(life_exp)
#| warning: false
#| min-lines: 6
#| max-lines: 8
#| fig-width: 9
wdi |>
  mutate(region = fct_reorder(region, life_exp)) |>
  ggplot(mapping = aes(y = life_exp, x = region)) +
  geom_boxplot() +
  scale_x_discrete(labels = label_wrap_gen(width = 15)) +
  labs(
    x = NULL,
    y = "Life expectancy",
    title = "Distribution of life expectancy, by region"
  )

Nigel Holmes and “chartjunk”

Chartjunk

Graphical decorations that do not improve the viewer’s understanding of the data

  • Grid lines
  • Big Pointless Objects
  • Artistic flair

Nigel Holmes

  • Graphic designer
  • Former graphics director for Time magazine
  • Distinctive “visual flair” in his work
  • Frequently criticized by Tufte for his elaborate use of “chartjunk”

Examples of Holmes’ work

Useful junk?

Run an experiment!

Image credit: Useful junk?: the effects of visual embellishment on comprehension and memorability of charts

Experimental protocol

  • Compared chartjunk versions of graphs to standard/minimalist versions of graphs
  • Tested individuals on chart description and recall
  • 20 subjects split into short and long-term recall groups

Measures

  • Response scores - did the individual correctly read/interpret the chart?
  • Preferences - which type of chart did the individual prefer? Standard or embellished?
  • Gaze data - where did the subject look during the experiment? At data regions or embellishment regions?

Results

  • Chartjunk did not improve or hinder comprehension
  • Chartjunk did not improve or hinder immediate recall
  • Chartjunk slightly improved long-term recall
  • Participants preferred the Holmes charts
    • More attractive
    • More enjoyable
    • Easiest and fastest to remember

Rethinking Tufte’s definition of visual excellence

  • Most of Tufte’s claims lack any evidence to support them
  • Pursuit of an objective truth - is truth absolute?
  • Chartjunk can be eye-catching and memorable (maybe even beautiful?)

Wrap up

Recap

  • Edward Tufte’s view on graphical excellence - “Above all else show the data”
  • Maximizing the data-ink ratio leads to potentially clear, effective charts for academics
  • “Chartjunk” can be eye-catching and memorable
  • The usefulness of chartjunk is debateable