We picked through 200 years of inequality charts and maps. In one sense, we are still grappling with the same challenges as Charles Dupin: understanding and comparing distributions. As we saw, the struggle is rich with new visual forms and analytic surprises.

You may also enjoy my recent associated essay on the topic:

Onward!—RJ

Subscribe now

About

Chartography is the newsletter of Visionary Press and Info We Trust, by RJ Andrews.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

Imagine you are watching a strange parade of people march by. Each person is sized by the money they make. A person of average height has an average income. Someone a little shorter than average makes a little less. Someone a little taller than average makes a little more. The first people in the parade are invisible, they are all in debt. Then, tiny folks no bigger than matchsticks flutter by. The parade continues, each marcher slightly taller than the last. Each marcher slightly richer than the last. The last person is a man of inconceivable height, several miles tall.

This parade vision is how Dutch economist Jan Pen gave readers a visceral introduction to economic inequality data in his 1971 book Income Distribution. Pen’s parade is modern and playful. But the impulse behind it, to make inequality visible, is much older. Over the past two centuries, we’ve developed a suite of ways to measure and visualize inequality.

Data graphics are machines for comparison. They let us see two groups side by side and ask: why are they different? The visual inquiry of inequality is a magical tool that has led to understanding and progress of many sorts.

Today, Wednesday February 25th, I’m talking live with , who writes the popular newsletter. His work looks beyond income and wealth inequalities to chart other ways America is unevenly distributed. Join us for our discussion here:

https://open.substack.com/live-stream/116379

To prepare for my chat with Jeremy, I reviewed the history of visualizing inequality.

This year we celebrate the bicentennial anniversary of Charles Dupin’s education map calling for the expansion of schools. What follows is a whirlwind tour of my favorite graphic gestures at understanding and conveying inequality since Dupin’s choropleth map, 200 years ago.

Moral statistics in space

Charles Dupin’s map (below left) is shaded by student density, expressed as a productivity measure: how much population is needed to “supply” a male pupil. Lighter districts have a higher proportion of students. He used the map and its accompanying tables to argue that education is the foundation of scientific and industrial power, education produces wealth, and education is correlated with a positive financial return. With Dupin’s map, administrative space became a statistical surface.

Early inequality visualizations were similar to Dupin’s map: French districts shaded in grays to show a variety of newly published state data. They were part of a movement called moral statistics sparked by André-Michel Guerry and championed by Adolphe Quetelet (1830s examples from both above). They advanced the radical idea that social phenomena obey laws, that society could be studied with the same scientific rigor as physics.

Moral statistics was particularly concerned with correlations between poverty and health outcomes—an association which still fascinates us.

By 1900, choropleth maps had traveled beyond France. They became colorful and more precise, with a late-century burst of vibrant examples tracking economic data at the building level: London poverty by Charles Booth (1889), Chicago wages by Hull-House (1895), and Philadelphia social-classes by W.E.B. Du Bois (1899)—all detailed below.

Economist curves

Maps are pretty. They are also powerful declarations of the breadth of available data. But it’s hard to claim precise assessments of inequality across hundreds of marks on a map.

For these kinds of behavior insights, inequality data must be compressed—abstracted—into summary values. Graphically, we also followed suit. Instead of making marks where people lived we piled them into buckets, stacked them in columns, and plotted abstract markers of populations as line-graph curves.

Maps show where inequality lives. Curves show how inequality behaves, across a population and across time.

Many of these statistical inventions survive in the modern analytic imagination, including: Pareto’s power-law distribution of income distribution (1896, below left) and Max Lorenz’s curves expressing the inequality of income and wealth distributions by comparing them to the 45° angle of equality (1905, below right).

Later, Corrado Gini reduced the gap between Lorenz’s curve and equality to a single number (1914), further compressing and abstracting inequality measures.

As tax systems and household surveys expanded, so did the data available to compress into curves. With new data and new statistical tools, nations and NGOs produced serious charts across the twentieth century of all sorts of inequalities. You can see these new ways of showing economic inequality in action in Wesley Mitchell’s Income in the United States (1921, below).

Another example of this genre is a 1938 package of charts by the U.S. National Resource Committee. I like how, together, they show that anchoring your buckets to different dimensions changes the analytic lens. Below, see the same data presented by income level, aggregate income, and families:

Selling inequality to the public

Post-WW1 corporate and government policy charts were often austerely-styled. But all color was not lost.

The 1930s and 1940s also saw a boom in persuasive (and propaganda) graphics in magazines and trade books for the masses. They used data graphics to attract attention and advocate for a cause. The finest examples of this genre were designed by Otto Neurath’s Isotype Institute. For example, see the below inequality charts from There’s Work for All (1945).

I particularly love chart VI (left), which gives 1% marks to both wealth and people, especially contrasted against the more conventional income towers of chart IX (right), which remind me of Pen’s Parade. See hundreds (and growing!) Isotype examples in my new explorer: https://isotype.infowetrust.com

Inequality goes digital

The Internet amplified all of these approaches to visualizing inequality. In my lifetime, there’s been an explosion of books and policy documents littered with global and national inequality charts. Some earned cute nicknames: Alan Krueger’s “Gatsby curve” shows higher income inequality associated with lower intergenerational mobility (2012, left). The Lakner-Milanovic “elephant curve” shows the unequal distribution of income growth for individuals belonging to different income groups (2013, right).

YouTube has given us a new medium for data storytelling: the mass-consumed short video.

In living memory, Hans Rosling used this new tool to best effect, painting dramatic stories of economic and health progress using digital morphing magic (left) and channeling Isotype with simple props that represent large aggregate values (right: each box represents 1 billion people across an income spectrum).

In 2012, Politizane published a viral video that dramatized the extreme differences between how we think about inequality compared to actual inequality:

More extravagant charts were developed too. Below is a stereogram set by Jack Blundell showing income deciles between 1980 (top) and 2014 (bottom). Each set is colored by the country’s 1980 position, which highlights extreme jumps in rank (see China). The richest decile for each country is in the back.

Early inequality graphics obsessed over poverty. Today, we obsess over billionaires.

This is expressed graphically by charting the rising dominance of the top 1%. For example, see the increasing share of national income by the top 1% (red) compared to the bottom 50% (blue) in the 2018 World Inequality Report.

I hope you are able to join Jeremy and me as we discuss these and other examples of visualizing inequality. https://open.substack.com/live-stream/116379

The graphic comparison at the heart of inequality—us versus them, rich and poor, haves and have-nots—feels older than money itself. The forms will keep changing. The comparison will not.

Onward!—RJ

Attention San Francisco!

I will be speaking about data-graphic masterpieces this Saturday at ABAA California International Antiquarian Book Fair at Pier 27. Details: https://www.abaa.org/event/speaker-schedule

About

Chartography is the newsletter of Visionary Press and Info We Trust, by RJ Andrews.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

I’ve been cooking—building an original archive explorer from my collection of Isotype Institute charts and maps, so I can find fresh insights and inspiration in seconds.

Now I can search charts the way my designer brain wants: by chart type, features, and color. I’ve also loaded it with publication metadata and OCR text.

You can play with Isotype Explorer too. Right here:

Isotype Explorer

So far, it includes 129 figures from 10 different publications. Here’s what the user interface looks like:

Isotype is a favorite source of inspiration for me and many information designers. For the uninitiated:

Isotype shows data with pictorial symbols. It consists of standardized methods and abstracted pictograms to represent social-scientific data.

Isotype is associated with Otto Neurath and his collaborators. It was first known as the Vienna Method of Pictorial Statistics, originating in the 1920s at the Gesellschafts und Wirtschaftsmuseum in Wien (Social and Economic Museum of Vienna). The term Isotype was applied to the method in the 1930s, after its key practitioners were forced to leave Vienna amid the rise of Austrian fascism.

This effort is inspired by:

• David Rumsey—digitization and open access

• Letterform Archive—gorgeous archival UI

• Jason Forrest—shared Isotype enthusiasm

• E-commerce—filtering & sorting that respects intent

Isotype Explorer has a few neat views. The grid view is great for discovery. Here it is with random sort, which creates unexpected juxtapositions.

The publication view helps you understand the total catalog. It’s shown below demonstrating chart and color filters.

The figure lightbox (below) details a single figure. Open it to read publication context, plus original caption text when relevant.

I find searching by a specific chart type to be really useful. Try: filter Type → Map → Choropleth (or just search “choropleth”).

The search bar queries publication and figure metadata, plus OCR text. You can also search by token keys. It currently supports: type, work (= publication), year, feature, and series. I’ve also added a British-to-American dictionary because these are mostly British publications. The tool understands theater and theatre, color and colour.

Despite these books being within arm’s reach for years, I never noticed some of their brilliant figures. For example, here is "Burden of Total Taxation.” I love how it compresses stacks of dots horizontally while maintaining the same diverging deciles.

Here is another surprise to me, “How Syphilis spread to the Old World.” It’s a perfect balance and density of arrows and symbols.

The whole site tracks state in the URL, so it’s easy to share any chart or filtered view. For example:

https://isotype.infowetrust.com/#/?sort=oldest&view=publications&id=w0002-p0088-f99

And it’s all original, including new photography, color correction, metadata, and website design by me.

I have many more publications, and hundreds of figures, to add.

You can help by playing around at isotype.infowetrust.com and let me know what you think.

• Reply with the coolest chart you find.

• Tell me what filter you wish existed.

• Flag any bugs you run into.

I gave some attention to mobile design, but it’s really best on your biggest screen. Enjoy.

Onward!—RJ

Isotype Explorer

About

Chartography is the newsletter of Visionary Press and Info We Trust, by RJ Andrews.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

Read more

When was the last time a new map really sparked joy in your life?

I am the proud recipient of “Daona-Youjiang: Echoes of the Rice Fields.” It’s a spectacular scroll map by the Guangxi Zhuang Autonomous Region Institute of Mapping. It’s over two meters wide.

Geographically, the map centers on China’s Youjiang River Basin in Guangxi province. Thematically, the map elevates the Zhuang people’s rice cultivation traditions, focusing on 那 (Nà) culture—“sweet” fertile lands.

To understand why this map exists, you need to meet the Zhuang people.

The Zhuang are the largest ethnic minority in China. Most live in southern China, especially in Guangxi Zhuang Autonomous Region, where they have lived for thousands of years. They have their own language, traditions, and festivals.

You can see Zhuang textile inspirations in the map’s indigo background and brocade framing. Zhuang clothing is known for its deep indigo dyes and bold geometric patterns, often accented with silver jewelry. Traditional Zhuang brocade—handwoven on wooden looms—includes patterns that encode stories, nature, and good fortune.

The map blends modern GIS techniques with traditional visual identities to depict Zhuang culture.

White waterways thread through bright yellow rice paddies—whose areas are derived from aerial imagery.

那 (Nà) is a phonetic loan-character from Mandarin used to represent ‘sweet’ or fertile land in the Zhuang language. Many places in Guangxi, such as the city of Nanning, incorporate the 那 character in their name. Scores of place names that incorporate 那 across the river basin are labeled on the map.

The geographic map is augmented with hand-drawn depictions of agricultural practices and festival customs of the Zhuang.

The way the map balances color is sublime. I particularly love its ocean of blues and pops of magenta.

I received the map over a live virtual exchange with its designers, who are Zhuang. In return, I sent them books from Visionary Press. Our common goal was to learn about each other’s craft, gain new inspiration, and build harmonious bridges across the Pacific. Let’s call it design diplomacy.

Despite so much distance, I felt a wonderful kindred spirit talking maps with designers dedicated to information-graphic excellence. I’m impressed by people using map-design so successfully to attract attention to their cause: promoting Zhuang culture.

They told me that the map is an expression of their feelings, emotions, and respect for Zhuang art and culture.

The team is, from left to right:
▪ Lan Jing, art designer
▪ Lan Rumen, map designer
▪ Hu Weibin, head of cultural and creative industry
▪ He Jiaheng, text editor

From our time together, I learned that in addition to indigo cloth, the map’s color scheme is inspired by the multi-colored glutinous/sticky rice of the Zhuang people.

And, if that wasn’t enough rice: The overall shape of the map represents the outline of a single grain of rice. (Scroll back up, you’ll see it.)

Part of the map’s design is its superior production. This map is not a JPG image file. It is a seven-foot scroll of lush paper, printed with rich inks, mounted on wooden rollers. Could any JPG image hold a candle to such a performance?

Production matters. Your canvas matters. Packaging matters. On that last note, the map arrived in a custom cloth sleeve that sported many Zhuang hallmarks: silver bangles, geometric brocade, and indigo blue. I’m considering fabricating a wooden box for it for long-term protection.

In-person matters too. The map is not a quiet object. The map is a bridge to new exchanges, new experiences, and new people.

The map gave me cause to connect with its inspiring designers, thousands of miles away. Today, it’s reason for us to gather in San Francisco, where it has already gone on display to information-designers in my studio, and Chinese-Americans at a local school. Come visit me in Dogpatch and you can enjoy it in-person too.

Before I go, one important note for paid subscribers: Soon I will send another newsletter inviting you to preview my latest project. It relates to Isotype Institute designs. I think you’re going to love it.

Onward!—RJ

About

Chartography is the newsletter of Visionary Press and Info We Trust, by RJ Andrews.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

Spurred by critique of his first design, Luigi Perozzo released an augmented version of his stereogram in 1881—adding “isotomic” (quantile) lines and publishing the plate with Latin inscriptions for international circulation.

It is with this second version that Perozzo adopts the word stereogramma.

As a refresher, the first design used a skeuomorphic volume to chart Swedish male cohorts over history. It used four colors, all employed again in his second design:

🟥 Red is years, from 1750 (left) to 1875 (right). Every 25 years gets a thick line.

⬜️ Gray is age, from birth (background) to 100-years-old (foreground).

🟦 Blue traces cohorts, everyone born in the same period.

🟩 Green marks constant population counts, isolines.

[Catch up with part 1 of this essay to better understand Perozzo’s 3-D project.]

In addition to being published in Latin, because he was sending copies to other national statistical offices, Perozzo’s second design boasted two new colors: golden yellow and orange. They are his attempt to make his stereogram more useful.

The new colors mark population quantiles. Perozzo called these lines “isotomic cuts”—cuts that divide a population into equal chunks. Eight quantiles, for example, divide a population into eight equal chunks.

Perozzo included two kinds of cuts, labeled in Latin: CENSITORUM SECTRICES (census cuts, yellow) and SUPERSTITUM SECTRICES (survivor/cohort cuts, orange).

🟨 Yellow lines slice a red year’s population into equal chunks.

🟧 Orange lines slice a blue cohort’s population into equal chunks.

The orange lines are relatively short because there are only a few cohorts in the data born early enough to reach age 100 by 1875.

Perozzo’s colors are engaging. But are they useful? It’s hard to say with historic Swedish data. Let’s try something closer to now.

I applied Perozzo’s method to the United States, illustrating its population since 1900, below.

Note that this is total population, not birth and survivorship curves. It encompasses many phenomena including birthrate and immigration.

I’ve never seen my own country in such meticulous detail. Here’s what I notice (annotated below):

🅐 First hump, which begins around 1920: high births (spurred partially by parents from the largest immigration wave in early 1900s) and improving child survival rate (water filtration and chlorination).

🅑 Silent generation valley: low births during Great Depression and WW2, plus low immigration after restrictive 1924 Immigration Act.

🅒 Post-WW2 baby boom, a big ridge that persists through time

🅓 Gen-X valley, and echo of the silent generation valley

🅔 Echo ridge: baby boomer kids

🅕 Fewer baby-boom grandchildren

🅖 Noticeable rise in 100+ year-olds by 2025

From one perspective, the story isn’t about the ridges. It’s about that Depression/WW2/immigrant-exclusion valley. That valley contrasts surrounding terrain to make them look like booms. That valley echoes forward with a smaller Gen-X population.

The yellow lines cut the solid into quantiles, each with equal population.

I find that the stereogram’s perspective and overall rise in population make these quantiles hard to read. For example, the yellow line in the foreground advances forward in age in the data, starting well below fifty-years-old in 1900 and ending above fifty by 2025. However, it optically moves backwards in the graphic.

This prompted me to return to 2-D to look at two slices from the stereogram, the 1900- and 2025 population profiles. These are the left and right walls of the stereogram, below.

Empty white lines cut into each year to make quartiles. Visually you can get a sense that each chunk represents the same equal population.

Remember, each is showing the character of the population in a given year, not the survivorship of a particular birth cohort. Seeing 1900 and 2025 like this advance three critical observations (beyond the fact that the second shape is much bigger because our total population is bigger):

• The shapes of the two populations are very different. 1900 is a triangle, loaded with kids and trailing off rather constantly. It’s a young country. 2025’s shape is more boxy, like a rectangle.

• There is a decade-plus quartile shift. The median age (middle white line) shifts from early 20s to late 30s. In 1900, 25% of the population were kids younger than 11!

• Reading the top ridge of 2025, detailed below, shows us that named generations are more than just marketing speak. Unique peaks and valleys suggest that our population composition will keep changing over time—and with it the composition of our interests and needs.

The USA stereogram differs in one significant way from Perozzo’s Sweden.

USA charts all population in a given year while Perozzo charted the survivorship of birth cohorts. The nice thing about his dataset is that the resulting cohort line always decreases. You might know this as monotonic.

A monotonic curve works great in 3-D because monotonic survivorship avoids overhangs, reversals, and other weird self‑intersections that inhibit views. They are more “terrain-like.”

USA’s full population is a different story. Its bumpy ridges block background valleys. The ridges also create difficult saddle-shape cells and ugly sawtooth transitions. For me, it was technically more challenging to render USA than recreating Sweden. Most critically, a ridge obscures populations behind it and makes it difficult to understand the full shape of the stereogram.

While they are the most enduring, Swedish stereograms weren’t the only fabulous vision that Perozzo pursued. Today we could render USA as a spinning object. We could slice and dice it digitally. We could even 3-D print it. But for now, I’ll just do what Perozzo did and pair the front view with a back view.

Having the back view shows us that, aside from the Gen-X valley, the original front view wasn’t missing too much (even though it felt like it was). The back view also reveals the Newborn curve and the age-profile of 1900.

There’s more to be gained from reviving sophisticated volumes. I look forward to developing more stereograms with other types of data.

Onward!—RJ

p.s. Paid subscribers get a bonus behind-the-scenes look at how I visually debugged the USA stereogram. Keep scrolling to the bottom of this newsletter to see.

For your radar

Three timely updates:

Nightingale talk + discount

On January 29th I am delivering a talk at the Florence Nightingale museum, “God’s Revenge Upon Murder” about Nightingale’s mortality diagrams.

Watch on Zoom or attend the London watch party at the Nightingale museum. Free, but registration is required.

Nightingale talk details

Plus, I’ve discounted my Nightingale book in honor of this talk. Order one from Visionary Press.

Principles of Data Graphics

I’m returning to Cooper Union to teach a five-session workshop starting April 21st, all virtual—but very alive and very interactive. We always have a blast with this class, it usually sells out so don’t dawdle.

Workshop Details

Freshy-fresh website

A nasty performance bug sucked me into a complete re-design of my online presence: avatars, images, copy, pages—you name it and I improved it.

This includes a new logo and color kit for Chartography (did you notice?) and a totally refreshed portfolio website. Check it out:

infowetrust.com

About

Chartography is the newsletter of Visionary Press and Info We Trust, by RJ Andrews.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

Read more

There’s a data graphic that has long-captured imaginations.

Its first version looks like this:

I’ve long marveled at this graphic, without really understanding how it works or why it came to be. Aesthetic appeal held my admiration—it’s like something from a steampunk future that never was.

They called this chart type a stereogram, named from the Greek stereos, solid, for its voluminous heft. They also considered calling it a triagram, recognizing that diagram references two dimensions, but rejected this as an impractical neologism. Today, you might know “stereogram” from Magic Eye optical illusions, a slightly different usage.

Over the past months, I’ve chased down (and acquired some of) its variants, read several contemporary papers about its development, and reconstructed the stereogram as an interactive digital version. I’ll show you the interactive in a moment.

Perozzo’s dream

The stereogram shows Sweden’s male live births and survivors by age class over time.

It was published by Luigi Perozzo, a twenty-four-year-old Italian engineer and statistician, in an 1880 volume of Annali di Statistica.

Perozzo chose Sweden because it boasted impressive national population statistics: centralized, continuous, frequent, and age-specific censuses since 1749 collected by the Church of Sweden. It was the very best data set from which to build a time-sequenced 3-D surface.

Why does seeing a population matter? If you manage a state, insurance company, or pension, then it is critical to understand demographic statistics. Perozzo created the diagram chasing two goals, which are often in graphic tension:

1. Represent statistics as clearly and precisely as possible.

2. Elevate analytic insights such as summaries, variation, and uncertainty.

Perozzo was convinced his stereogram could accomplish both goals simultaneously.

Graphically, Perozzo was inspired by contour and relief representations at the 1878 Congrès international de démographie in Paris. In particular, he singled out an 1860s chart by Fredrik Berg, the first head of Sweden’s Statistiska central-byråns.

Mathematically, the stereogram-approach built on the work of a group of German visionaries including Georg Knapp, Karl Becker, Wilhelm Lexis, and especially Gustav Zeuner. To get his particular 3-D projection, Perozzo blended Zeuner’s three-axis approach with a 60° tilt inspired by Hungarian Jakab Lewin.

As part of Perozzo’s skeuomorphic perspective development, he also seemed to explore the data by fabricating physical models—but the extant evidence of this is thin.

Perozzo’s design research and development helped him recognize that a 3-D object brings clarity to particular features of a population and makes its overall character more memorable. He wanted to make population feel more real and the stereogram was how to do it.

How it works

To detail Perozzo’s stereogram, I’m going to use my digital rendition, which allows us to establish it in layers. Then we’ll see what it has to say.

Start with a 2-D grid, below.

🟥 The red horizontal dimension is years, from 1750 to 1875. Every 25 years gets a thick line. Every five years gets a thin line.

⬜️ The gray dimension is age, also marked in 25- and 5-year increments, from birth to 100-years-old in the foreground. (This project taught me that quinquennial is the term for every five years.)

These are the x- and y-axes. This grid is the base of the stereogram.

Next, raise every intersection in the grid to a height representing the population size at that point. For example, the gold ring below marks how many men were in the 25-year-old age bucket in the year 1775. (That bucket includes all men ages 21-25).

With points meshed into a 3-D surface, add shading to reinforce the stereogram’s dimensionality.

On top of the 3-D surface, Perozzo layered two analytic colors.

🟦 Blue (below left) traces cohort survivorship. See, for example, how many people were born in 1750 and then follow the line downward until the few survivors of that cohort are 100-years-old in 1850.

🟩 Green (below right) are isolines marking constant values. They help us read how many people exist at each point. For example, you can read the increase in 25-year-olds: 100,000 in 1795 and 150,000 in 1850.

Put all the quinquennial red-gray-blue-green lines together and you get a rather stunning object. Here is Perozzo’s 1880 original again:

At a glance, what can we see?

• A progressive, albeit noisy, increase in number of births. The final 50 years is a population boom. This is emphasized by the large swoopy Linea delle Nascite.

• A harrowing child mortality rate. Each blue line begins with a horrible fall down a steep cliff, from birth to age 10.

• By age 100, nearly everyone is dead.

Then, look a little deeper, sprinkle in a little history, and you will see more.

The stereogram reveals that shocks to birth—war or disease—don’t just show up once. They propagate diagonally as the cohort ages. See how any initial dip in births triggers a diagonal valley that continues through the ages.

These valleys have echoes too. As a smaller cohort reaches adulthood, they produce relatively fewer children, resulting in a new valley. See how births take a dip around 1810, likely because of the Napoleonic wars. Births take another dip 30 years later when the war-thinned cohort reaches family-forming age. Since they are fewer in number, they have fewer children. And then, 30 years later, there is another dip around 1870!

I masked the original stereogram to emphasize these valleys:

The same war echoes for generations. These demographic waves illustrate long-lag consequences, something that Perozzo emphasized would be difficult to appreciate without the stereogram.

Luigi rides again

I transcribed one of Perozzo’s data tables and recreated the stereogram in SVG. It’s now a digital interactive, with a simple tooltip that teaches you how to read the colors:

Digitizing the stereogram taught me how difficult it is to simulate a solid object in a convincing way—a way that balances all design elements. It also exposed how small changes in perspective, color, and shading can totally change the emphasis of the construction.

This project benefited from encouragement by Tony Chu and Santiago Ortiz. You can play with the interactive here:

see interactive

Admiring masterpieces like Perozzo’s stereogram, I have often wondered why we can’t achieve similar lavish levels of design digitally.

The reason is that it is really really hard. The whole thing is an illusion: a pile of 2-D lines and shapes and text, all data-driven, that come together in a convincing way to fool us into thinking it is 3-D. Creating a simulated 3-D object from real demographic tables is hard. Rendering a surface, architectural elements, and labels in the same 3-D coordinate system is hard.

The boon is that I now have a reusable engine, and can do it again, quicker. I can now turn other data sets into 3-D illusions.

A success?

Perhaps you are still not sold on the stereogram. You would be in good company.

Wilhelm Lexis, pioneer of the analysis of demographic time series, translated Perozzo’s paper into German the same year. He followed the translation with his own observation, acknowledging that while three-axis representation is scientifically rigorous, it still may hinder usability.

Because demographic quantities can be related in so many ways, a stereometric (3-D) construction, having one more dimension, can in itself achieve a greater completeness of immediate representation than a planimetric (2-D) one; but these 3-D constructions are more complicated, harder to survey at a glance, and therefore more inconvenient for practical use than the 2-D ones; … so in practice the planimetric constructions will always keep the upper hand.

In response to Lexis, and determined to prove that the stereogram could be useful to practical analysis, Perozzo refined and elaborated his design. The resulting graphic, version two, was published the following year of 1881. It has been lauded by generations of information designers since.

We’ll study Perozzo’s refined vision, and expand his methods to modern visual problems, in a future edition of Chartography.

Onward!—RJ

Update: the second part is here:

Two timely notices for your radar this week:

Nightingale talk + discount

On January 29th I am delivering a talk at the Florence Nightingale museum, “God’s Revenge Upon Murder” about Nightingale’s mortality diagrams.

Watch on Zoom or attend the London watch party at the Nightingale museum. Free, but registration is required.

Nightingale talk details

Plus, I’ve discounted my Nightingale book in honor of this talk. Order one from Visionary Press.

Freshy-fresh website

A nasty performance bug sucked me into a complete re-design of my online presence: avatars, images, copy, pages—you name it and I improved it.

This includes a new logo and color kit for Chartography (did you notice?) and a totally refreshed portfolio website. Check it out:

infowetrust.com

About

Chartography is the newsletter of Visionary Press and Info We Trust, by RJ Andrews.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

The past year of Chartography spanned the philosophy of design, practical know-how, and a lot of inspiration from the history of data graphics. We also launched my new book Info We Trust Remastered and the Maps for Kids poster series.

By the numbers, Chartography was read more than 150,000 times in 2025. That includes the attention of over 1,000 new subscribers. Everyone will benefit from a fresh look at one or more of the top five essays from the past year. Here they are, ordered according to my own capricious whim.

One of 2025’s first essays was a real heater. You ate it up. I still don’t say “data visualization” (except when it gets me paid) and have noticed that “data graphics” is catching on.

I enjoy information philosophy and adore historic charts, but practice is the heart of it all. For any of my theory or research to be valid, it must connect to how I make new charts that have to be really useful. Sometimes it’s a challenge to abstract lessons from my client practice. Not so with “Color Engineering”—it’s an approach that begged to be codified and shared.

It took me a summer of reading and research to come to grips with clarity—a quality often extolled as the ideal of information design. It never sat right with me. Digging into its true meaning revealed why.

A near-acquisition led me to discover the most violent story behind a new chart form. It’s a perfect example of how charts help us claw our way out of disaster. (That acquisition? It happened. But the volume arrived missing its key fold-out chart. The hunt continues.)

I had a theory a few years ago that Nazis, with their mechanical mania and insidious indoctrination, probably made some interesting charts. Since then we’ve surfaced many striking examples worth study. and I examined one collection:

As I review these essays, I’m proud of the variety of stories told. I’m also grateful for the quality of their writing. Writing helps me interrogate the world, make sense of it, and crystallize understanding. I couldn’t sink effort into these if I didn’t know you’d be reading. Thank you for your patronage—your attention, your paid subscriptions, your reply emails, your in-person compliments. They mean a lot.

Writing from the United States, 2025 was a year of upheaval. For those dedicated to craft and advancing our understanding of reality, it wasn’t the most harmonious trip around the sun.

One of the songs I’ve returned to this year is Stevie Wonder’s “As”—for its sermon of a bridge:

We all know sometimes life’s hates and troubles
Can make you wish you were born in another time and space

But you can bet your life times that and twice its double
That God knew exactly where he wanted you to be placed

So make sure when you say you’re in it, but not of it
You’re not helping to make this earth a place sometimes called Hell

Change your words into truth and then change that truth into love
And maybe our children’s grandchildren
And their great-grandchildren will tell
I’ll be loving you

I choose to be cheerful, productive, and strive to engage in as many in-person connections as possible. Looking back, it surprises me how much that attitude helped me be productive, and hopefully bring some good into the world:

• Hosted monthly lunches at my new library/studio in Dogpatch. The amount of design talent we have in one room frequently feels dangerous. Reach out if you are in San Francisco and want to come to the next one.

• Published Info We Trust, Remastered—my love letter to data storytelling, with a book-launch party at Letterform Archive.

• Curated and mounted The Craft of Data Graphics exhibition in New York, elevating the work of living and historic information designers.

• Edited the Clarity summer series, republishing classic essays about information design.

• Two-dozen historic publications collected for my design library. (Come visit it in SF!)

• Launched the new Maps for Kids poster series and sent posters to kids all over. Discounts and free shipping are still available for a short while!

Chartography will return in 2026.

Onward!—RJ

About

Chartography is the newsletter of Visionary Press and Info We Trust, by RJ Andrews.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

Most world maps skip the main character: life. They carefully mark countries, cities, and borders, but not the plants and animals that make Earth unique. We think that’s foolish.

Earth Biomes is an illustrated wall map, part of the new Maps For Kids poster series. Across three views of our globe, colors mark different biomes—rainforests, deserts, tundras, and more—while ocean blues darken to show where the water runs deep.

The impetus for this design was the need for a traditional geographic map as part of the inaugural Maps for Kids trio. It complements the other two types, a pictorial diagram (Steam Locomotive) and a statistical chart (Food Energy). Together, they make a rich information-graphics demonstration.

This map is a collaboration between us—RJ Andrews and Robert Simmon. Today, we’re going to reveal how we wrestled overwrought boundaries, honed our layout, and finessed delicate color balances. The explanations below are neatly organized. The work itself wasn’t: it was messy and iterative. Our process only emerged by making the thing.

Layout

One of the final steps of our technical pipeline, layout composition, was the first serious design consideration tackled: How do we arrange our globes?

Our poster is inspired by classic double-planisphere maps, which pair two hemispheres side-by-side to display the whole Earth.

For example, we really enjoyed several spreads by Erwin Raisz, such as The Living Planet from his Atlas of Global Geography (1944).

After many attempts, though, we realized that there is no good way to split the Earth into two halves and retain all major landforms unbroken. You really need three views to appreciate major features uninterrupted. Below are early composition attempts, blurred to emphasize visual mass.

These didn’t work. The geometric arrangement on the left feels dead—and its Mickey Mouse ears are distracting. The middle one gives us bigger globes by only displaying a portion of one, but we really need three full views to include New Zealand (whose often omission has Kiwis justifiably upset). The organic flow on the right is too messy.

Our final triangular arrangement is cocked slightly counter-clockwise. This gives dynamic tension; it is more lively than the original triangle’s regularity but not as messy as the organic flow.

Landforms overlap where our three globes nearly touch. To emphasize the southern globe’s flipped relationship to the top two, its text is upside-down.

Data

We started with the World Wildlife Fund’s major habitat types, part of its terrestrial ecosystems dataset. These define 15 biome types (including ice), each with a rather technical name.

Considering our audience, kids and their grown-ups, we combined some biome types, reducing the total to 11. We renamed most. For example, “Tropical & Subtropical Moist Broadleaf Forests” became “Tropical Rainforests”.

You can see the effects of this refactoring in the before and after views of South America. Notice the large grassy region on the right, which includes the Pampas. It was made by combining two original biome types that were hard to distinguish to our audience, seen on the left:

• Temperate grasslands, savannas and shrublands (green)

• Tropical and subtropical grasslands, savannas and shrublands (lavender)

Sea-depth bathymetry data was sourced from the GEBCO 2024 Grid. More on that soon.

Representation

Precision is a tricky aspect of maps. Too much detail looks fussy, and can result in distracting artifacts. Too little precision and important features may be distorted or disappear altogether. Ideally, the resolution of the data matches the map’s scale and purpose.

One approach to reducing detail is to remove some of the points that form the edges of features, but this tends to create jagged, artificial boundaries. An alternative solution is to smooth edges. Often, we use a combination of the two techniques.

Smoothing is simple in principle, but can be difficult in practice. Small features tend to disappear—often the very details that make a map intriguing. There’s a balance between enough smoothing to look good, but not so much that it removes important information. The below animation shows before and after-smoothing view of the Gulf of Mexico.

Notice how the coastlines and biome boundaries keep their character while reducing details that would only be distracting noise at our printed poster’s scale.

It required several smoothing attempts to preserve the “Flooded Grasslands and Savannas” covering the southern tip of Florida.

Our bathymetry data also required transformation. We downsampled, smoothed using a Gaussian blur, and quantized into regular contours.

Color

As usual with a map, choosing colors was one of the most challenging design decisions. In this case, we needed to balance discernibility—whether the individual colors stand out from each other—with readability—whether the colors make sense.

Each biome needed to be easy to distinguish from its neighbors, and be recognizable even without reading the legend. Typical qualitative color palettes used for differentiating discrete categories are often garish, filled with highly saturated primary colors.

Maps For Kids is largely inspired by Isotype Institute designs — pictorial statistics for the masses with flat colors and bold symbols. Our biome hues (and degree of smoothing) take direct notes from a 1930 map, Vegetationszonen der Erde, by Otto Neurath.

We chose to use a similar palette, with desaturated Earth-tone hues grouped into similar categories (different types of forest are different shades of green, woodlands and savannas incorporate more yellow, and desert is nearly beige). This choice, combined with merging related biomes, helps reduce visual noise while keeping each biome distinct. It also strikes a fair balance between appearing natural without coming across as a strict physical map. Ben Day patterns (tiny dots floating above the solid colors) further distinguish various forest biomes.

The other issue with color was balancing the biomes with the much larger ocean area. A single ocean fill-color would have emphasized the land, but at the expense of being rather bland. We chose to illustrate low-contrast bathymetry, aggregated into 10 levels with a nonlinear color scale.

Our ocean color ramp is wedged into a narrow value-band that maintains contrast with the ice biome’s white and the background’s cream. We created it in Adobe Illustrator using a custom script that interpolated colors between low and high values.

Inspired by Neurath, we also attempted warm oceans—see below—but determined them too avant garde for an audience that expects water to be blue.

We also considered a dark-background variant, as if the globes were floating in space. Alas, too planetarium.

Our ocean “blue” is in fact shifted subtly toward green to complement the cream-colored paper. The result helps the land areas stand out while giving interesting texture to the large ocean areas, especially in the Southern Hemisphere view.

Text and other annotations

One feature that makes a map a map is its annotations: the text that turns marks on paper into information. In the legend, small icons show the dominant vegetation for each region. Simple gauges indicate temperature and rainfall—two of the key ingredients that define biomes.

On the map itself, we overlaid a faint graticule, giving the globes a sense of volume and quiet authority.

Labeling the biomes was trickier. Many sprawl across multiple countries and languages; some have no common English name at all. We chose to label only the largest biomes, taking cues from the marine and terrestrial physical features in Natural Earth and adapting them for a primarily American audience. Each biome label is set small and in low contrast—you can read it when you go looking, but it mostly disappears when you take in the poster as a whole.

Our small block of introductory text gives just enough for a grown-up to to begin exploring the map with a kid.

Production

We discussed every design detail together, collaborating especially closely on color. Robert led the cartography, handling datasets, transformation, and projection. RJ led illustration.

Afterward, we made the below diagram to capture the cascade of tools, exports, and processes necessary to create our map. In practice, this flow emerged organically over several months and was never run all at once.

Both Bay-Area residents, we are lucky to have collaborated on this several times in person, reviewing print samples along the way. We also negotiated dozens of hiccups and decisions over iMessage. This particular exchange represents an ideal of creative collaboration: positive enthusiasm while engaging with the nitty-gritty.

Is the design a success? Adults who’ve seen it in person say it’s beautiful.

But the real test is in front of a child, unrolled on the kitchen table, hung in a bedroom, or stuck on the fridge. Does it catch their attention? Does it prompt a question? Does it spark their understanding and imagination?

As we write this, the first Earth Biomes posters are on their way to kids across the globe.

If you’d like to put the living planet at arm’s reach in your own home, and see how it performs in your own ecosystem, you can order one here:

Order Earth Biomes

About

Chartography is by RJ Andrews and the newsletter of Visionary Press and Info We Trust.

Robert Simmon is a cartographer and science communicator who specializes in visualizing satellite imagery. His work work has appeared on the front page of the New York Times and the cover of National Geographic. He is also known for creating the first global view of the Earth at Night, and crafting the Blue Marble featured on the original Apple iPhone. See his work at RobertSimmon.com.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

This is my favorite time of year: I get to send beautiful information graphics all over the world (over 60 countries and counting!).

This holiday season, I’m joyfully celebrating the new edition of Info We Trust, the launch of MAPS FOR KIDSS—and doing something I’ve never done before:

• the biggest discounts on books and prints

• two very limited bundles

• AND Visionary Press is giving free domestic (USA) shipping on everything.

Shop Visionary Press Holiday Deals

2025 has been an extraordinary year for Visionary Press.

Earlier this year I released Info We Trust, Remastered with events in San Francisco and New York. It’s a rebuilt edition of my book about the craft and value of data graphics. The text is totally rewritten, many illustrations are redrawn, and the whole object is a big upgrade: printing and binding that combine to a beautiful-to-hold package. This is where I’ve poured everything I know.

MAPS FOR KIDS, the new poster series, starts from a simple belief: almost anything can be treated as a map. The first three posters are now in the shop—Earth Biomes, Food Energy, and Steam Locomotive—each a big, richly illustrated explanation designed to be enjoyed by kids and the grown-ups reading with them. 

They make wonderful gifts.

One thing I’ve noticed is that every few weeks someone orders “one of everything” from Visionary Press. It’s one of the biggest votes of confidence I can get. Inspired by these enthusiasts, I’ve created two limited holiday bundles:

Visionary Press Library: four books that map the landscape of information graphics.
– Info We Trust (Remastered)
– The three Information Graphic Visionaries volumes

Visionary Press Collection: A full bounty of information graphics:
– 4 books (the full Visionary Press Library)
– 3 MAPS FOR KIDS posters
– 2 large foldout prints: Cross-Sections Through California and Neil & Buzz

If you’ve been waiting for the right moment to add these to your studio, classroom, or a kid’s bedroom wall, this is it. The discounts are real, quantities of the full collection are very limited (only a dozen!), and every order ships free within the U.S.

Shop Visionary Press Holiday Deals

Thank you for helping me put thoughtful and delightful information graphics into the world.

Onward!—RJ

About

Chartography is by RJ Andrews and the newsletter of Visionary Press and Info We Trust.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

MAPS FOR KIDS is here.

Over the past year you’ve watched this project take shape: early topic lists, logo sketches, research trips, Ben-Day experiments, naked diagrams waiting for labels. Today the first three finished wall maps are real objects in the world—ink on thick Japanese matte paper—and you can finally order them from Visionary Press.

Shop Maps For Kids Posters

Maps for Kids starts from a simple idea: almost anything can be treated as a map. Geographic projections, statistical charts, and pictorial diagrams are all ways of flattening the world so we can see and understand better.

But the real magic happens when we enjoy those maps together. Each design is meant to be enjoyed side by side—loaded with kid-sized points of engagement and quiet cues for parents, big siblings, and teachers to help narrate what’s going on.

Let’s meet our three maps.

1. Food Energy

A chart about how everyday foods power our bodies.

Food Energy arranges familiar snacks and ingredients into a colorful wall chart. 146 foundation foods are placed by how much of their energy comes from fats, proteins, or carbohydrates, so kids can literally see the difference between an avocado, a yogurt, and a bowl of rice.

Learn what each macronutrient does — quick energy, steady energy, growth energy. A small inset focuses on carbohydrate-rich foods and shows which ones get most of their energy from sugars. The result is a poster that works as both a kitchen conversation starter and a gentle statistics lesson. Kids compare what they love (and don’t love) to eat while grownups finally have a map to all the foods they’ve eaten their whole life.

Hang it on the fridge, in a classroom, or anywhere people enjoy food.

2. Earth Biomes

A portrait of the living planet at arm’s length.

“Earth Biomes” is an illustrated wall map built from three globes. They line up so continents continue across them, with the bottom globe flipped to reveal the South Pole. Color washes across the land to mark major biomes—rainforests, deserts, tundras, and more—while ocean blues darken to show where the water runs deep.

This design is a collaboration with Robert Simmon, earth-visualizer extraordinaire.

A legend of plant silhouettes and tiny “recipes” of thermometers and raindrops shows how temperature and rainfall combine to shape each landscape. As kids explore, patterns snap into place: green belts hugging the equator, deserts hiding in mountain rain shadows, icy crowns at the poles. It’s an invitation to see Earth not just as “land and water,” but as a set of living systems.

I’ve shown its test prints to many people and each one had the same reaction: That’s beautiful!

3. Steam Locomotive

A diagram that shows how heat turns into motion.

“Steam Locomotive” tells the story of a classic wood-burning 4-4-0 engine—the kind that helped stitch the United States together in the 19th century. The engine stretches across the page like a mechanical animal, every major system illustrated and explained.

A short guiding narrative walks you from the tender full of wood and water, into the firebox and boiler, up to the steam dome, and finally down to the piston and driving wheels that pull the train forward. Around that journey spot extra details: the whistle that sings with steam, sand that helps slipping wheels grip the rails, and a tall smokestack that catches sparks before they set the prairie on fire.

After using it to explain how a train works to my 3-year-old, I was startled to find him soon drawing cutaway diagrams of steamies, revealing their firebox and boiler tubes inside.

How to get them

• Buy a single poster for $45

• Or get all three for $99

The three-poster package is a balanced introduction to information graphics: one diagram, one chart, one map. Steam, snacks, and the whole planet—a tiny gallery of how we can map different kinds of knowledge.

$45 Choose One

$99 three-poster set

What they’re printed on

These are not flimsy handouts.

Each poster is printed on thick, Japanese-made matte art paper—the kind that feels substantial in your hands and doesn’t glare under bright lights. Colors are crisp, so fine labels and little details pop.

Posters are rolled in protective tissue and shipped in a triangular carton, which means they arrive ready to gift: no creases and no tube wrestling

Behind the scenes, an international network of printers produces the posters close to where you live. That reduces shipping distance and customs friction, so your order tends to arrive faster (and cheaper) than if everything had to travel from California.

For everyone in the United States, Visionary Press is offering free domestic shipping for Black Friday and Cyber Monday.

Who these are for

• Kids who ask “How does that work?” and “What’s that called?”

• Grown-up enthusiasts for inspiring design

• Folks like teachers, librarians, and pediatricians looking for visual explanations that respect kids’ intelligence

• Everyone hunting for a holiday gift that isn’t another plastic thing with batteries

I appreciate you following my year-long series chronicling this passion project. Thank you for cheering Maps For Kids along while it was still wobbly. This is the moment when the sketches and stories turn into artifacts you can actually unroll, hang, and enjoy.

Order Maps For Kids

I’m excited to finally send these maps out into the world—and onto the walls of the kids (and kids at heart) in your life.

Onward!—RJ

About

Chartography is by RJ Andrews and the newsletter of Visionary Press and Info We Trust.

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

On a Sunday in May 1842, the world’s faith in progress went up in flames outside Paris.

A special excursion train was returning from the King’s annual fête at Versailles. The train—two engines followed by seventeen wooden carriages—was packed with nearly 800 partiers.

Near Bellevue, a locomotive axle snapped on the descent to the city. The iron horse pitched over the rails, the boilers ruptured, and the front passenger cars jackknifed.

In moments the wreck became a furnace.

It gets worse. The car doors had been sealed—perhaps to prevent fare-dodgers hitching a free ride, perhaps to prevent passengers from injuring themselves while the train was in motion. After the crash, some passengers escaped through the windows. Others, injured or pinned in the wreckage, were unable to move. Trapped inside, they burned alive. Their screams haunt several eyewitness accounts.

More than fifty people died, including Antarctic explorer Jules Dumont d’Urville and his family. The Versailles-railway catastrophe was not just a mechanical failure, and not just the deadliest railway accident ever. It was an early reckoning with the speed and scale of industrial life. The accident revealed how thin the line was between harnessing energy and being consumed by it.

Could the speed of progress be tamed?

In the months that followed, engineers picked through the twisted metal to analyze what had gone wrong. One of those engineers was Jules Petiet.

Petiet’s Diagram

Next spring, Jules Petiet’s technical analysis of the accident appeared in Journal des chemins de fer (5 June 1843).

Petiet was an impressive fellow. He would soon be Chief Engineer of the Chemins de fer du Nord [Northern Railway Company]. He then designed locomotives before leading the prestigious engineering school École Centrale Paris. Petiet’s name is one of 72 inscribed on the Eiffel Tower.

Around the same time as his analysis came out, Petiet also published a new type of diagram that promised to help avert future rail disasters. Tracé géométrique de la marche et de la composition des trains (“Geometric layout of train movement and composition”).

Petiet charted movement over time, showing train paths snaking down stations across the course of the day. His diagram may be lost to history, but a contemporary description detailed that it also encoded the importance of each train, as measured by number of carriages, in the train’s line thickness.

Descendants of Petiet’s time-distance map are still created today. See a modern Caltrain example below.

Each red or blue string represents a train, some move up the train-line and some move opposite, down. Circles highlight where red express trains overtake blue local trains. The diagram helps identify pathing conflicts (string intersections), train speeds (string slopes), and station dwell times (horizontal strings).

To our passenger’s eye, the Caltrain diagram is a fascinating novelty. Curious, but not as helpful to catch a train as a simple timetable. Petiet’s diagram, however, was not invented for passengers.

Ibry’s Diagram

Petiet’s original graphic method was immediately put into use by Jacques Alphonse Busche on northern railway lines. It was also picked up by Charles Ibry, deputy chief operating officer of the railway from Paris to Rouen. (On some documents, Ibry is spelled Ybry.)

In 1847, as part of a promotion blitz to deploy the design cross-country, Ibry presented the design system to the Minister of Public Works. Its summary report included the following plate showing the Paris-Rouen line.

Like the Caltrain diagram, stations go down the page and time goes across it. Notice the criss-crossing trains, land elevation to the left of the station names, and dark horizontal bands.

The dark horizontal bands represent tunnels. For example, the top band occurs just above Bonnières. It’s the 2,600m (1.6 miles) Tunnel de Rolleboise, which opened in 1843. You can see it on this period map of the Paris-Rouen line as a dashed line.

The other feature of this plate is its skeuomorphic angled ruler. It is an astonishingly instructive detail. In 2019, design historian Sandra Rendgen gushed for Ibry’s machine (emphasis hers):

what is MOST EXCITING is that the article features a graphic showing a hardware instrument for plotting the schedules– this piece does not simply show a diagram, it shows a board with a flexible ruler screwed onto a little plate which can glide along the bottom of the diagram. By adapting the angle of the ruler, the engineers were able to schedule an individual speed for each train. This chart is absolutely unique – it makes us understand that these charts were not just meant to display a pre-conceived schedule—no! The charts and the instrument were actually used to set up the schedule in the first place.

The speed-indicating needle (H) at the bottom of the sliding ruler (B’) is geared (DEF) to turn with the ruler’s angle. This enabled an engineer to fix the ruler’s slope to a particular speed and then plot the train’s path.

Look at the slope from the elevation plan, and schedule the train’s speed accordingly. This graphic method does not just visualize schedules. It made the schedules.

Ibry’s 1847 campaign was a success. Railway diagrams swept the contintent, and were then expanded to other modes of transportation such as omnibus.

The Bibliothèque nationale de France holds two early blank diagrams, one with Ibry’s name on the bottom. They are labeled with the diagram’s name: RÉGULATEUR DES CHEMINS DE FER. Railway regulator. These extant templates are still waiting to be inserted into the machine and have their trains plotted.

Decades later, Léon Lalanne reflected that there wasn’t a single railway timetable that is not the numerical translation of graphical constructions “previously designed to ensure, for a fixed number of trains determined a priori, the most economical, safest, and most advantageous operation for both passenger and freight service.”

Marey’s Diagram

Étienne-Jules Marey honored Ibry’s diagram by including it in his La Méthode Graphique (1878, 1885), which I recently translated with Georges Hattab (Visionary Press).

Marey’s cartoon rendition kept station names and time of day but left off the elevation plan. (He also [mis]attributes the invention to Ibry.) Marey admired that “the line corresponding to each train expresses: departure and arrival times, relative and absolute train speeds, time of passage at each station, and duration of stops.”

By the time Edward Tufte used Marey’s graphic for the cover of The Visual Display of Quantitative Information (1983), it was completely detached from historic context. On Tufte’s cover, station names and times are absent. Inside, Tufte credits only Marey. The Versailles wreck, Petiet’s origin, Ibry’s promotion, and everything else was forgotten. The chart had become pretty furniture—a billboard for the book that introduced a generation to data graphics.

To recount:
1843 — Petiet charts it.
1847 — Ibry mechanizes it.
1878 — Marey abstracts it.
1983 — Tufte immortalizes it.

Howard G. Funkhouser included Ibry’s campaign in “Historical development of the graphical representation of statistical data” (1937), referencing the 1847 report. That’s how I knew it existed, and how I found and purchased the volume containing that remarkable plate showing Ibry’s machine in action.

Research, guided by a sense that an extraordinary thing might be out there, is a pillar of rare works collecting. It’s one of its joys. In a way, this essay is a tidy version of what went through my mind as I hunted for that early graphic. (Not included: my many dead ends.) I am interested in sharing more about collecting and understanding old charts.

I am also interested in how data graphics come to be. They are born mostly from necessity—too often from horrific necessity. Time and again, charts have helped us claw our way out of disaster. From tragedy, a data graphic of modern transport was born. How many lives did it save? We thankfully will never know.

Onward!—RJ

About

RJ Andrews is obsessed with data graphics. He helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

Attila Bátorfy: Do you remember when and why, or perhaps who, sparked your interest in the history of data graphics?

RJ Andrews: I’m the son of two history nuts (and collectors). Nearly every vacation was to a museum, often to go behind the scenes with curators into collections. In September 2016, I visited Tableau in Seattle to give a talk to their research staff. Afterward, I went for a walk with Ben Jones, leader of Tableau Public at the time, around Gas Works Park. Ben encouraged me to look into the history of data visualization, observing that no one from my generation had a good handle on it. The following fifteen months was my swift turn into historic graphics. Ben’s prompt led me to create my Ogilby timeline, which is mostly inspired by Michael Friendly’s Milestones project. The project was my first attempt at understanding our tradition. Its original poster got a lot of attention, which encouraged me to turn it into an interactive and keep digging into historic charts.

In July 2017 I pitched a publisher on a book about the history of charts titled Conveyed; how data revealed the world. They didn’t spring for that concept, instead funding the original edition of Info We Trust.

That autumn, I attended the first Ruderman Conference at Stanford’s David Rumsey Map Center. There, David Rumsey introduced himself to me, citing his enthusiasm for the Ogilby project. He confided that he was turning his attention toward collecting historic data graphics and invited me to his house to see his collection.

There, in November 2017, he told me that I was about to start collecting. He was right. The following month, inspired by encountering real works in David’s collection, I made a batch of facsimile linen-back case maps as Christmas presents for information-graphic friends. My accompanying card neatly summarized the past year’s journey:

Enclosed please find a handmade and to-scale facsimile of Darton & Gardner’s Comparative Heights of the Principal Mountains and Lengths of the Principal Rivers. The famous and earliest known rivers-and-mountains type graphic, it launched decades of similar charts. The 1823 original was produced as a “case map”: fixed to a linen backing & contained within a cardboard sleeve. When I first saw one of the rare existing originals in person I was amazed at how small it was and thought you might be delighted with a similar experience! It is packed with whimsy and design flourishes that are sure to inspire. 2017 has been a wonderful year of learning for me and I am so thankful that you have been part of the adventure.

Looking back, this was obviously a specter of my future publishing endeavour, Visionary Press.

What do you think is the reason for that the history of data graphics/information graphics became popular in the past decade?

It doesn’t feel popular to me. Any rise in the past decade may be due to newly digitized archives, data journalism’s rise bringing attention to the field, plus a now-extinct vibrant international community on Twitter.

Thanks for reading Attila Bátorfy’s Cabinet of Infographic Curiosities! Subscribe for free to receive new posts and support my work.

Which period do you find the most exciting and interesting?

Works surrounding the first world war appeal to me. Before the Great War you had the zenith of print culture and empire powers, plus the rapid rise of statistical science. This combination produced monumental achievements, including lavish American and French statistical atlases. After the Great War, prior achievements diffused to new places where they were spun around with fresh visual approaches to voice strongly persuasive arguments, or propaganda across the world. My prior example would be Scaife’s Civil War, and later examples are the Istanbul Journal of Justice and the Isotype Institute.

How does the inspiration you draw from old data graphics appear in your information design work?

The history of data graphics teaches us two essential things. When we make data graphics we are part of a proud tradition of people all over the world and all over society who worked hard to better discern reality and create shared understanding. It’s never been easy, but it’s often inspiring and mostly worthwhile. Studying information-design history also teaches us more ways to solve information problems. Like us, our predecessors faced all sorts of information problems. The solutions they naturally arrived at are often different from how we would solve things. This is because they existed with different contexts: analog workflows, big canvases with high resolutions, large creative teams, different data, and different narrative pressures.

If you look at the literature on history to date, do you have any general observations about the works?

My Information Graphic Visionaries book series put forward the kind of history I am interested in: one that gets into the details about what, how, and why masterpieces are made. We are ready to move past histories that are too superficial, especially long lists without connective tissue between important works. I also have a sense that no one has struck the right geographic balance in their telling. Previous histories are too western-centric, true. They didn’t properly acknowledge universal and ancient aspects of information graphics. But modern tellings reach for some fantasy where the bar chart, and other artifacts of statistical science’s development, is as old as time and equally distributed everywhere. And that’s just not how it happened. These are tools that were invented and developed by a relatively small number of people—adopted, adapted, and improved over time. That story is interesting to me.

Are there any data graphics that you specifically discovered and introduced to the public? Please tell us about them!

I noticed, researched, and collected Frances Lightfoot’s colossal chronographer. I am also the first to organize the catalog of Florence Nightingale’s graphics.

Have you ever thought about what will happen to your collection when you are no longer around?

I’ve assembled a collection of data graphics that I consider a working reference design library—a resource for me and others to help us make great work. It’s a great joy to host tours and expose people to great works. I have an original Minard design treatise from 1861 that is quite precious. It inspired me to translate its accompanying text. There are logical ends for any collection. It can be sold for parts, transferred to an institution for preservation, or passed on to an heir (Or some combination.). I began collecting at a relatively young age, and believe I still have some time to figure things out!

What was the reception of the first three volumes of Visionary?

Beautiful books are very important to me. Like others before me (Nightingale, Brinton, Tufte), I embarked on a financially foolish endeavour to make beautiful data-graphic books. I wish I didn’t have to and that a traditional publisher would do it for me. Let me know if you are that publisher and can take over the burden! Making nice books is really hard. Reception and enthusiasm was enough to put the project into production, but not enough to sell out our print run. Get your books before they sell out. I don’t expect to ever be able to reprint these at such a phenomenal quality.

Do you have any new book plans for the Visionary series?

I’m working on a poster series titled Maps for Kids inspired by classic Marie Neurath books. It takes the perspective that everything is a map. We will launch with a map of an American steam locomotive, a map of Earth biomes, and a map of macronutrient energy composition across two hundred basic foods. I want to use this series to develop beautiful pictorial diagrams, statistical charts, and geographic maps.

The five greatest data graphics and/or creators of all time? One cannot be American, German, French, or British!

Acknowledging that these dudes often, or mostly had other people actually produce their designs, my top five today are: Halley, Playfair, Humboldt, Walker, and the Neuraths. Each developed several novel information graphics with lasting impact. And, Neurath was Austrian, not German ;)

Subscribe to Attila Bátorfy’s Cabinet of Infographic Curiosities” newsletter.

About

RJ Andrews helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

I’m designing a Maps for Kids poster that plots 200 foods by macronutrient composition. Each mark is a tiny picture—so I need hundreds of symbols that read at small sizes, keep stylistic cohesion, and still feel exciting.

Here’s my hand-to-digital system that makes that possible.

Style Inspiration

Isotype is my primary reference for Maps for Kids. Their symbol style, crafted by Gerd Arntz, features hand-cut silhouettes—bold, chunky masses with nicks, wobble, and tiny printing artifacts that give them a tactile feel. Proportions are slightly volumetric, in profile or ¾ view.

Here’s a selection of Isotype and Isotype-inspired food icons from around the 1930s I assembled as reference:

Compared to modern Noun Project-style icons, Isotype is less sterile and less geometric, with varied line weights, irregular edges, and specific, narrative detail that conveys materiality and category at a glance.

Design Constraints

I’m not trying to imitate Isotype pictographs exactly. My needs and goals are a little different:

1. Show hundreds of different symbols at once (Isotype only used a few in one composition)

2. Each symbol is printed relatively small

3. Each symbol can have more than one color to help differentiate them and make the poster more exciting

The Isotype qualities I aspire to are roughness, handmade feel, and occasional ¾-perspective.

Workflow

After attempting a digital-only workflow, I developed a process that originates with pencil sketches, continues with marker fills, and then proceeds to Photoshop to separate and clean color layers before choosing specific colors. The easy trick to get a handmade feel continues to be … make it by hand.

Subscribe now

Here’s what the progress looks like for my watermelon symbol:

But a single symbol isn’t enough.

I need to do this for hundreds of foods in a way that maintains style while also differentiating each food. (Yeah, I briefly tried to use AI to do this. It couldn’t. I also considered stock image libraries.)

Here are the first pencil sketches. Each was made while looking at “studio lighting” photography of real food.

Below are the marker fills for the same illustrations.

You can see early attempts with blue and gray markers on the first page. I quickly learned that black-red-yellow markers separate nicely and stuck with them for the rest.

The marker fills are rough and full of wobbly lines. In a word, human. Perfect!

These illustrations are the primary ingredient to my digital workflow. But I’m not publishing a red asparagus—my icons need another ingredient: color.

Color System

Isotype symbols rarely had more than one color. Marie Neurath described in Instructural Science 3 (1974):

Colours are also part of the visual language. Their meaning is often traditional, but often we had to introduce our own traditions. We tried not to use too many colours with too many meanings, as pictures should be read easily and also remembered if possible; too many colours would confuse. … We tried to adhere more and more to the rule that even in black and white a chart could still make sense.

I developed a color palette based on historic Isotype colors. Its swatches are below: Columns are hue families; rows are value from dark (top) to light (bottom). Background swatch approximates the poster’s field.

Why so many greens? I need several options to help detail all the leafy vegetables.

Having a consistent color palette helps me move quickly—I never have to think about which red I should use. It also helps the overall composition hang together because all of the icons are colorfully associated.

Subscribe now

Coloring is done with masked color fill layers in Photoshop. This let’s me quickly eyedrop colors from my palette. Notice the layered use of masks.

Photoshop detail: I made fill layers from the scanned marker illustrations by: selecting color range, and then modifying selection with smooth (2px), expand (2px), and adjust levels to get rid of any color ghosts.

Toolchain

The color palette is a living document (Illustrator) that live links to the icons (Photoshop) which live links to the poster composition (InDesign). Here’s what my system looks like:

And here are the icons so far. Not quite 200, but we are getting there!

I’m developing these as quickly as I know how, and it still feels slow. Perhaps not artistically slow, but economically slow.

Is this level of effort worth it? Looking forward it hardly ever does. It seems like there’s a waterfall of content out there more quickly produced that will garner much more attention. But looking backward at a dozen years of nifty projects, I’m glad I always put in the work.

Critique

Seen together, I spot a few symbols that may need more love: that spinach leaf needs a second color, I don’t like the heavy lines on the juice box, the onion needs another try.

But overall I’m pleased with the result. It’s not Isotype, it’s its own thing. And that’s good.

The symbols also do not have to do all the lifting: on the composition, the icons will be accompanied by labels too. I’m still playing with label size and color.

I’m also considering distinguishing the labels for any foods that get more than half their energy from a special carbohydrate: sugar. Perhaps by changing their label’s color?

By now, you may have noticed that my symbols lack protein and fat. That’s because they are nearly all carb-heavy foods. Let’s see how they all look laid out on the carbohydrate corner of the ternary diagram.

Read more

I’m struggling to make a new color palette for a project.

So I built a small tool based on how I think about color design: Color can be engineered in a repeatable way that clarifies messy choices.

Color is hard

A simple way to understand meaning is as a game of difference. We don’t care as much about absolute values as we do relative contrasts. Young and old, rich and poor, kind and cruel. Things often matter because we compare them to something else.

Knowledge deserving more emphasis in your brain gets more graphic emphasis on my canvas, often by marrying semantic contrast with graphic contrast. Left and right, big and small, light and dark.

Color is a magical phenomenon. Its properties beautifully match our need to differentiate information. From the color chapter in Info We Trust Remastered (Visionary Press, $39):

Distinct colors draw our attention because we perceive color according to difference. From a survival perspective, colors are most relevant in relation to their environment. Over the course of a day, the quality of light is changing as the sun moves through the sky—both in brightness and in color—but we seldom perceive this. Our pupils physically adapt to the changes in lighting by contracting in bright conditions and dilating when it is dim. Meanwhile, other parts of our visual system compensate for the shift from yellow-white light when the sun is high above to the orange and red light when the sun is low on the horizon. . . .

Imagine you are hunting for ripe red apples among a green expanse of leaves. To contrast the red color of desire against other environmental cues, our eyes and brains adjust to changing conditions. This adaptation lets us hunt for fruit at noon and dusk, on an open plain or in a forest. We change to make effective comparisons across various lighting conditions.

Color’s interdependent nature also makes it hard to nail the right color palette. Sometimes I find the right mix intuitively, almost artistically. Other times I turn to a more mechanical method that anchors on the idea of using color’s ability to differentiate, contrast, and emphasize.

Two things to know

There are lots of ways to model color. A popular method comes stock with digital design tools, HSB: Hue accounts for named colors around the color wheel. Saturation is color intensity or strength—how far away something is from gray. Brightness is related to how dark or light a color looks. (This luminance dimension is sometimes presented as lightness or value.)

For example, for a fixed blue hue, see how color changes as brightness increases upwards and saturation increases rightward. Try starting with the center dot, and move your eyes up (brighter) or down (darker) or left (more gray) or right (more vivid). Then, look at the corners to see combinations (e.g. upper right corner is brighter and more saturated).

If colors are to help us differentiate information, they must be well-differentiated.

What’s curious about color is that brightness is generally a stronger differentiator than hue or saturation. This has physiological roots in our visual system: we are wired to notice light–dark edges first. Equal-brightness hues tend to fuse at small sizes or in the periphery. Even a modest step in brightness separates marks cleanly.

Designing with brightness first is also more resilient to color deficiencies, which may cause confusion between two colors distinguished only by hue.

Color is a lifetime study; but for information design, two points can do the lion’s share of the work: (1) Color can be broken into three dimensions. (2) Color’s brightness dimension is the most useful to differentiate.

Brightness → hue → saturation

This sequence reflects how color can help differentiate and emphasize information:

• First, set brightness levels. The color that deserves the most emphasis gets the highest contrast compared to the background color. Colors that should be well-differentiated should have different levels of brightness. Aim for at least 10 points in your tool’s brightness/value scale between neighbors; thin lines or small labels may need more.

• Next, choose your hues. Beware semantic opportunities (water is blue) and traps (red may not be best for positive connotation).

• Finally, adjust saturation. This step achieves aesthetic coherence—it helps everything hang together.

Let’s explore this method by analyzing some of the colors in the below map detail of Takamatsu’s public transit by Jug Cerovic.

I used my eye dropper tool and found the numerical value of brightness, saturation, and hue. Then I plotted them together:

There are two groups of colors: transit-line marks and basemap. The groups are differentiated from each other by both brightness and saturation dimensions: see how each group exists in their own vertical bands of brightness and saturation, which I’ve highlighted with yellow and gray background fills.

Brightness and saturation distinguish each group. Then, inside either group, colors are differentiated by hue. See how each color owns a unique part of its hue circle.

The first time I made this color-deconstruction plot, a lot of the mercurial aspects of color snapped into mechanical focus. It also helps me explain color design to others. (This approach originated once upon a time when a client begged for a technical explanation of how a proposed palette worked.)

Extracting color dimensions, plotting their values, and iterating design is a pretty clunky workflow. Since my first plot, I’ve only reached for it to teach color design and when I’m in a real pickle. Until now.

Color tooling in practice

I made a simple color tool that plots swatches, their component dimensions, states their contrast ratio compared to the background color, and lets me adjust component values. Here it is analyzing the new NYC subway map. You can see a saturation distinction between background and transit groups. (Note that I’m using Value here instead of Brightness.)

The tool gives me a perspective on how color works. It’s also useful beyond analysis, for actual iterative creation. This method gives me hints on where to nudge colors toward more effective palettes. For example, here’s a horizontal bar chart that needs a strong hierarchy: a “hero” color (dark blue) that I want to distinguish from secondary and tertiary categories:

Analyzing the palette, I like that the hero color’s value (brightness) is far separate from the others (red circle). But the overall effect still has unwanted clash.

Diagnosis: Some color attributes may be too close. Background and 2nd blue have similar brightness value. Hero and 2nd blue share similar hue and saturation. Both are opportunities for further distinction.

Remedy: Chang the 2nd blue, dramatically altering its hue and saturation. Modify the 3rd color (gray) to put the brightness values and contrast ratios of the Hero-2nd-3rd sequence in the proper order. Here’s the palette remade addressing those unwanted neighbors:

To my eye, these colors have a stronger visual hierarchy. The overall effect is also better because the colors don’t compete as much for attention. It is more calm in a useful way.

You’re invited to play

Made in Google Sheets, this tool is super flexible. It can drag-and-drop rows, sort by brightness to spot near-neighbors, or add a new column that gives me the specific gap between neighboring brightness levels. (Under the hood it’s just a few HEX2DEC and DEC2HEX tricks.)

This tool is still a work in progress. Maybe someday it will evolve to something with a polished UI.

There is a link at the very bottom of the email for paying subscribers to play with the tool. Make a copy and have fun. Tell me where it soars and where it fails. (And if a paid subscription is too much of a burden for you right now, message me and I will get you access.)

Rolling your own tooling is fun. But you gotta build something with it! I still need that new color palette for my project.

Onward!—RJ

p.s. This is a companion to A Dangling Blue, my rant against cold inhuman design tools.

About

RJ Andrews helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

Read more

Lucid clarity conjures clearness: clear judgment, clear intellect, and clear conscience. Clearness of sight and clearness of sound.

English polymath Thomas Browne heard it in the “clarity of their understanding” (Pseudodoxia Epidemica, 1646). Millennia earlier, Aristotle emphasized that the proper function of style is to make the meaning clear (Rhetoric III-2). That’s lucid clarity.

Designers long called this duty perspicuity—transparent expression in the service of understanding. You may not know that word. (I didn’t.) See below how 🔴CLARITY’s popularity has risen in the wake of 🔵PERSPICUITY’s decline.

This switch helps us better appreciate the subtitle of William Playfair’s Statistical Breviary (1801): “Charts Representing Physical Powers … with Ease and Perspicuity.” Today, Playfair would say “with clarity.”

The modern and partial understanding of clarity, lucid clarity, is simply perspicuity. Metaphorically, think of lucid clarity as a polished window. It is the kind of frame that lets you see what’s going on uninhibited: clean, efficient, simple, and obvious (once you see it).

But clarity is never actually used to describe a window. When we grade something on its clarity we are admiring diamonds, not windows.

Lucid clarity’s triumph feels total. Yet something vital is missing.

Today, the demand for lucid clarity is ever louder as attentiveness and concentration shrivel under the pressure of digital media.

Lucid clarity’s insistence on easy insights makes no room for the intricacies of reality. Lucid clarity has little appreciation for astonishment, amazement, and wonder.

For example, lucid clarity has no way of acknowledging why I am attracted to fantastical historic attempts, like this detail from the U.S. Treasury Department (1897, BLR). This polar diagram offers little in the department of “clear meaning,” yet it still excites me to look:

One hazard of chasing lucid clarity is confusing presentation and understanding. Graphic efficiency is only loosely related to effective reading. Don’t mistake the messenger for the message.

The essential clarity, like all meaning and value, lives in the mind of the viewer. A clearly-expressed chart alone in the woods inspires no clarity.

Simple truths aren’t always best served by plain design. Here are two treatments of the same data—one lucid, one luminous.

In the lucid example below, a single line averages the story into one curve. It’s graphically efficient and makes a crisp claim.

This line graph was created by Randy Olson to spotlight attributes of the Wall Street Journal original:

Perhaps the lucid line graph makes a clearer argument. But it lacks that mysterious thing that the luminous original radiates with. The table texture is exciting, which makes its post-vaccine silence more powerful.

Simple truths are not always best conveyed through plain design. This point is sometimes made through the rebuke of analytic lenses like “data–to–ink ratio.” Lucid-clarity obsessives might have us refrain from the funfetti spectacle of the disease-incidence table. But when clarity is reduced to frictionless decoding, gratification of mind departs.

Our world feels awash in information windows but devoid of information diamonds.

Lucid clarity, modern clarity, is not wrong. But it is incomplete.

Lucid clarity makes charts easy to read—light as spotlight. Luminous clarity uses light as architecture. It makes things worth inhabiting.

If we look deeper into clarity’s history, we see these richer qualities: brightness, luster, radiance, splendor, and glory. Clarity is not merely clear thought. It is also spiritual illumination.

Greek philosopher Plotinus gives the luminous sense: clarity as radiance that seizes the knower. “What is it that attracts the eyes of those to whom a beautiful object is presented, and calls them, lures them, towards it, and fills them with joy at the sight?” (Enneads, c. AD 250)

Gabrielle Merite’s much-loved hug chart, above, fills me with that joy. From my perspective, it isn’t differentiated by its lucidity, but by its extravagant luminosity.

Here again, it is easy to confuse messenger and message. The glowing attributes of her composition and the luminous clarity it encourages in me are related, but not the same. The supreme glow is the one inside.

Plotinus described entering into union with knowledge as beautiful: thrilled with delight, stirred to harmonious coherence and affinity with splendor beyond material bodies. “This is the spirit that Beauty must ever induce, wonderment and a delicious trouble.”

The eyelids close; yet a light flashes before us. … All that one sees as a spectacle is still external; one must bring the vision within and see no longer in that mode of separation but as we know ourselves; thus a man filled with a god—possessed by Apollo or by one of the Muses—need no longer look outside for his vision of the divine being; it is but finding the strength to see divinity within.

Suddenly, swept beyond it all by the very crest of the wave of Intellect surging beneath, he is lifted and sees, never knowing how; the vision floods the eyes with light, but it is not a light showing some other object, the light is itself the vision. … Radiance.

Now, Plotinus was not specifically describing the sensation of reading a bar chart, but his words resonate with how I understand the magic of information graphics. Illumination does not stop at shining the light on something else, illumination swells when your own soul becomes one with the light it perceives.

Clarity is light worth beholding.

Knowing used to be godliness.

Think of knowledge as unification with a power beyond our corporeal form. This resonates with my excitement for the possibility of datasets: Knowledge lets us grasp something bigger than ourselves. Data lets us stretch across vast space and time, giving us vision to phenomena impossible for any single person to ever experience on their own.

German polymath Hildegard of Bingen provides a striking example of luminous clarity. She was a twelfth century abbess now credited as a founder of scientific natural history. She is also famous for visions, possibly the result of migraine headaches, which she described and illustrated. Read about one of her visions from Scivias (c.1151):

I saw a great mountain the color of iron, and enthroned on it One of such great glory that it blinded my sight. Before him, at the foot of the mountain, stood an image full of eyes on all sides, in which, because of those eyes, I could discern no human form. In front of this image stood another, a child wearing a tunic of subdued color but white shoes, upon whose head such glory descended from the One enthroned upon that mountain that I could not look at its face. But from the One who sat enthroned upon that mountain many living sparks sprang forth, which flew very sweetly around the images.

… He Who was enthroned upon that mountain cried out in a strong, loud voice saying, “O human, who are fragile dust of the earth and ashes of ashes! … Burst forth into a fountain of abundance and overflow with mystical knowledge, until they who now think you contemptible because of Eve’s transgression are stirred up by the flood of your irrigation. For you have received your profound insight not from humans, but from the lofty and tremendous Judge on high, where this calmness will shine strongly with glorious light among the shining ones.

And here is that vision illustrated. Note the blue figure, clothed in eyes—vigilant awe made visible.

Hildegard’s vision shows what luminous clarity feels like—a blazing union of seer and splendor. It also echoes the medieval account of phantasia (the inner “imagination”), later dramatized by Dante as images that “rained into [his] high imagination” (Purgatorio XVII, 25).

About a century after Hildegard, Thomas Aquinas named beauty’s three marks: integritas, consonantia, claritas: wholeness, proportion, and radiance.

Cautions

In his meditations, René Descartes further distinguished the separation of perception and knowledge, seer and truth:

I am so amazed that I can neither touch the bottom, nor swim at the top. … I am a Thinking Thing, that is to say, doubting, affirming, denying, understanding few things, ignorant of many things, willing, nilling, imagining also, and sensitive.

Whatever I clearly and distinctly perceive is certainly true. But I have formerly admitted many things as very certain and manifest, which I afterwards found to be doubtful.

Descartes cautions us against blind faith in clarity. Clear perception with our senses is not a guarantee of truth. “I know not, whether they are true or false, that is to say, whether the ideas I have of them are the ideas of things which really are, or are not.”

You may look with perfect vision to the night sky. But what you see is not perfectly true. Try again, this time with a telescope.

We last heard Aristotle endorsing perspicuity, the importance of clear meaning. He also adds an ethic about proportion of emphasis. Proper style requires the appropriate amount of splendor. Grandeur about trivialities rings false. Blinding radiance misleads, or worse: overpowers, oppresses, dominates.

Yet, I am wary we’ve lost too much illumination via too much mechanical efficiency.

Take care to not over-decorate, sure. But that is hardly a risk today. I feel lost in a Flatland where there’s a lot of everything and it all feels the same.

Italo Calvino touched on this in his short story “How much shall we bet?” (Cosmicomics, 1968), where an eternal being longs to be again at the beginning of creation:

And I think how beautiful it was then, through that void, to draw lines and parabolas, pick out the precise point, the intersection between space and time where the event would spring forth, undeniable in the prominence of its glow; whereas now events come flowing down without interruption, like cement being poured, one column next to the other, one within the other, separated by black and incongruous headlines, legible in many ways but intrinsically illegible, a doughy mass of events without form or direction, which surrounds, submerges, crushes all reasoning.

We are drowning in Calvino’s doughy mass of events. I yearn for the revitalization of his prominent glow.

Like past generations who reacted to industrialization by turning to the unity of light, body, and soul, I re-embrace clarity. I am now able to seek beyond clear thinking. Clarity is both lucid and luminous, both perspicuous and radiant. I seek the treasure that concludes the 1989 adventure film “Indiana Jones and the Last Crusade.”

Indiana Jones: What did you find, Dad?

Dr. Henry Jones, Sr.: Me? Illumination.

Onward!—RJ

Welcome to Chartography.net — insights and delights from the world of data storytelling.

This summer, we republished a series of historic writing about information design. It is the SUMMER OF CLARITY! These essay inspired the blue marginalia in my new book Info We Trust ($39 from Visionary Press).

About

RJ Andrews helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is currently out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

This summer, we are showcasing a series of historic writing about information design. It is the SUMMER OF CLARITY! These essay inspired the blue marginalia in my new book Info We Trust ($39 from Visionary Press).

Housekeeping: This is the final edition from our series republishing historic texts. But it is not the end to the Summer of Clarity. Here’s what I have been working on:

• A print keepsake inspired by this series for paying subscribers to Chartography, published by Visionary Press. Please consider a paid subscription today.

• A new essay by me reflecting on the concept of clarity, inspired by texts that predate modern statistical graphics. The digital discourse about clarity is all wrong, and I look forward to setting things right.

I look forward to presenting these in the coming weeks. Until then…

Today, please enjoy Georg von Mayr on data graphics, excerpted from his 1877 Die Gesetzmäßigkeit im Gesellschaftsleben [The laws of social life]. This is a case where the figures (seen in new photography) may be more innovative than the text (newly translated). Enjoy!

Ways of Representing Statistics

… In addition to numerals and words, however, the graphic representation of statistical results has become more and more popular in recent times. This serves primarily to popularize statistics and therefore deserves special consideration at this point.

For the time being, however, let us briefly consider the number and the word as a means of representing statistics. The number is the most original manifestation of statistics; it must be there before everything else, for without numbers there are no statistics in the modern sense of the word. The numbers themselves, however, do not appear in a colorful jumble, but in the well-ordered form of a table. The table presents the quantitative results of mass observation in a neat and clear grouping according to internal and external classifications. The tabular form is more than justifiably hated by the reading public. This is probably due to the fact that the table requires concentrated thinking, whereas the public prefers watered-down thinking.

It is obvious that the science of statistics cannot do without the word. If the word explanation is mentioned here at all as a “means of representation” of statistics, this is only due to the fact that in the past, many people, especially in the circles of official statisticians, were of the opinion that statistics should and should only produce [numeric] figures and nothing more. Today there is hardly anyone who seriously shares this limited view.

When official statistical tables were just beginning to be published, it was still possible to indulge in the delusion that private science would soon throw itself with particular enthusiasm on the [numeric] figures provided and utilize them in the most diverse ways. Time soon cured the official statisticians of this delusion, for it turned out that statistics, which only produced [numeric] figures, was a dead child. The statistician who conducted the survey and compiled the tables is first of all obliged to use his [numeric] figures scientifically and to criticize them himself. This is, of course, not possible without an explanation. If the statistician himself does not venture into the sea of [numeric] figures that he publishes, he should not be surprised if the layman, unfamiliar with the cliffs and shallows of this sea, shies away from it from the outset and pays no further attention to the series of figures presented to him.

Graphical representation requires more explanation than numbers and words.

The graphical method includes both the simple geometric illustrations of statistical figures—the diagrams—and the representation of statistical relationships on the map - the cartograms. These two types of graphical representation are essentially different from each other; it is therefore advisable to discuss them separately here.

I. Diagrams. The following can be considered for the geometric visualization of statistical figures: the point, the line, the area, the solid.

1. The point. The point as such has hardly any significance for pure diagrams. Since the point lacks any relationship of size, it cannot be used to represent anything other than the unit of the statistical numbers. Therefore, if a diagram were to be made absolutely in points, as many points would have to be made as the statistical numbers in question contain units. This would not provide a clear illustration even for small numbers, especially since the dot diagram would not require a specific grouping of the dots in a regular distribution on lines or areas. Rather, such an arrangement would indirectly produce a line or area diagram, and essentially only the technique of dotting would remain. However, this must not be confused with the dot diagram. The point is not used for quantitative visualization, but only as a means of distinguishing areas. Dotting here simply competes with hatching and color.

Incidentally, the uselessness of the point for statistical graphics only applies to “pure” diagrams. If the topographical representation is added to the diagram, in other words if the diagram and the cartogram are combined, then the point also gains significance for statistical graphics, as will be shown in more detail below.

2. The line. The line is used in diagrams in two ways, first as a straight line of different lengths (Fig. 1a), and secondly as a line connecting the end points of such straight lines (Fig. 1b).

The fact that in the first case only straight lines are chosen is recommended for clarity of representation. For the same reason, parallel straight lines constructed at right angles to a base line are chosen.

The following Figure 1 contains samples of such line diagrams.

Moreover, it cannot be denied that the method of representation, according to which straight lines of different sizes are placed next to each other without a connecting line, lacks clarity and has a downright disturbing effect on the eye. The eye wanders inactively between the end points of the straight lines, which are at different heights, and it is very difficult to miss the guiding connecting line between these end points. The situation is different if, instead of lines, surfaces that touch each other are lined up.

Here the eye finds, on the one hand, the necessary resting place through the width extension of the individual surfaces and, on the other hand, the desirable guidance through the graduated line formed by the contact of the different surfaces. In addition, the juxtaposition of the surfaces in front of the lines offers the great advantage that further distinctions between the overall quantities intended for representation can be made much more clearly and technically more easily. The line is much less easily distinguished by multiple hatchings and colors than the surface. Even with simple technical elaboration, this is evident, for example, from a comparison of Figure 2A and B.

According to this, the mere juxtaposition of straight lines of different lengths does not appear suitable as a line diagram. However, this is the case with the other type of line diagrams, which appear as the connecting lines of end points of different straight lines. Such line graphs, which are often, if not always with mathematical justification, called “curves”, have been used many times and in an appropriate way and will always remain an important tool of statistical graphics.

The advantage of this type of linear diagram lies above all in the fact that it offers the eye a simple and reliable guide in tracing the ascending and descending numerical relationships. In addition, they have the further significant advantage that they only take up a minimum of the area intended for the graphical representation and thus allow the addition of one or more comparison diagrams in the same linear representation. This is the case, for example, when mortality and birth rates are shown together with grain prices and temperature conditions in a diagram containing four comparative lines. Such comparative line diagrams, which can be made very easily distinguishable by coloring, dotting, breaking and other patterning of the lines, greatly facilitate the detection of any parallelism or antagonism between different factual phenomena. They are therefore of interest not only as illustrations of statistical figures for the consumers of statistics, but also to facilitate comparative research for the producers of statistics.

The line diagrams of the type just mentioned are preferably used where it is a question of illustrating simple members of statistical series of figures which contain no internally varying structure and move in regular sections, e.g. to show the price history of certain commodities, the annual movement in the number of deaths, crimes, etc.

A drawback of line diagrams is that there is no fixed internal relationship between the width and height of the overall representation. It is completely arbitrary how far the distance of the straight lines, which are chosen for representation by themselves or by connecting their end points, is taken.

The effective height of this straight line is just as arbitrary. In the first case, only the equality of the initially arbitrarily determined distance is required, and in the second case the proportionality of the initially likewise arbitrarily determined height. From this it follows that the same statistical series can be represented in a number of externally more or less dissimilar diagrams. If we pay particular attention to the diagrams shown in connecting lines, we can see how a “spiky” or a “flat” line diagram can be chosen with completely the same mathematical justification. Figure 3 gives an example of this.

In the previous discussion of line diagrams it has been assumed that a straight line forms the basis on which the lines proportional to the numbers are constructed perpendicularly. However, this is not necessary in itself; rather, any curve could be chosen as the base. Of course, the clarity of the representation would then suffer greatly, so that in practice the straight line will always retain the upper hand. Only another type of line, namely the circular line, can make some claim to consideration under special circumstances. The line diagram drawn in a circle can be used with some justification to depict relationships that actually represent a cycle. This is the case, for example, when mortality by month is not to be shown for a single calendar year, but according to the average of an annual series. In this case, January is in fact as close to December as it is to February, which can only be achieved by plotting in a circle, but not on the basis of a straight line. Fractions of the radians, measured from the center or from the periphery, represent the straight lines whose size is proportional to the numerical results. Extensions of the needles beyond the periphery can also be used for this purpose. The linear connection of the end points of the relevant straight lines is also useful here.

Examples of line diagrams in a circle are shown in Fig. 4 below. For a and c the center, for b and d the periphery forms the starting point of the straight lines, which represent the numerical differences of the statistical series as radius parts of different sizes. In the case of a and b, the illustration is only attempted by the different sizes of the radius parts. For e and d, the connecting line between the end points of these straight lines is added.

3. The area. The dot diagram can hardly be considered for statistical graphics and the line diagram, as can be seen from the previous presentation, is only permissible under certain assumptions. Accordingly, for all other statistical relationships, insofar as they can be usefully represented by diagrams at all, only the area diagram remains, since the solid diagram, as will be shown below, can only be used in only exceptional cases and is of no special significance for the statistical literature.

An essential advantage of the area diagram, as already indicated in the criticism of the line diagram, is that it permits the exact representation of the inner structure of the statistical relationships. However, the detail to be distinguished in the representation must not be too manifold, otherwise the same defect occurs here as with line diagrams, in which too many lines of comparison cross each other.

The conditions under which area diagrams appear appropriate are as follows.

Only simple figures may be chosen. For individual overall facts whose subdivisions are to be shown, it is advisable to use squares, which are divided into rectangles according to the subdivisions. Fig. 5 [above] shows a sample area diagram with two subdivisions.

If several statistical facts are to be compared graphically by means of surface representation, the use of rectangles appears to be the most suitable. For this purpose, rectangles with the same base and different heights or rectangles with different bases and the same heights can be chosen. The first of these methods of representation is the more descriptive, as can be seen from a glance at Fig. 6 [below]. It is closely related to the line diagram and can be seen everywhere, with particular advantage when it is necessary to visualize statistical subdivisions. This type of area diagram is therefore widely used in practice. It is desirable, and greatly facilitates the detailed study of these diagrams, if one square of the net, in which the upright rectangles are drawn, corresponds to the unit or a multiple of the statistical ratios shown, falling within the decimal system.

Figures other than rectangles can only be used in area diagrams in exceptional cases.

As far as the triangles are concerned, they could be used in series instead of the rectangles with the same base that follow one another, since they behave like these rectangles in terms of their area. However, a test (Fig. 7) shows that the rectangles are clearer and that the correct estimation of the height of the triangles is made somewhat more difficult by the size of the lines and the unequal angles.

However, the greatest reservations arise against the series of triangles when the representation of the inner structure of the statistical facts comes into consideration. While this is extremely simple and easy to understand in the case of rectangles, it requires extensive geometric calculations in the case of triangles and therefore makes the estimation of areas extremely difficult.

A large number of triangles in a row is not particularly suitable for comparative representation, but individual triangles can be used in a particular case to illustrate a peculiarly structured fact.

Such a case exists when it is important to clearly express a qualitatively important minority in relation to a massive majority. In particular, the triangle is the most suitable figure for the diagram when the individual components of the overall fact represent an ascending order, as is the case, for example, with the total number of educational establishments in a country, from elementary school up to universities.

By subdividing the triangle horizontally, the surfaces reach an ever-increasing height in relation to their content as they approach the spike, the statistical relationships shown at the apex become particularly clear without the proportionality of the surfaces being affected. If one were to represent the same thing in a square, the qualitatively significant but quantitatively small minority would almost completely disappear. By choosing the triangle, statistical graphics succeed in such special cases in achieving something more than merely converting numbers into areas. Fig. 8 [above], which shows the same composition of a complete fact in a triangle and in a square, serves as a model for comparison.

If rows of rectangles or triangles are chosen as diagrams, they are constructed on a straight line, usually a horizontal line. The polygon can only be used as a basis for concentric or outward-facing triangles in exceptional cases. This form of representation is internally justified in the same cases in which the line diagram can be selected in a circle. Examples of this are shown in Fig. 9.

It is not practical to use circular areas for the comparative representation of statistical relationships because the circle, despite its regularity, includes an area that is difficult to estimate. It seems more appropriate to subdivide the circle into circular sections to show the statistical division of a whole.

From a technical point of view, it should be noted that area diagrams allow the most extensive use of color and hatching to distinguish individual diagrams and their sub-areas. In particular, a careful combination of color and hatching makes it possible to express multiple relationships in one diagram. Care must be taken, however, that the diagram does not become artificial and incomprehensible.

4. The solid. If it is important to visualize statistical relationships in a rather rough way, the choice of the solid as a diagram can be useful, e.g. for exhibitions. For literature, however, this type of diagram is out of the question, since it is not possible to enclose boxes with the books, which are intended to popularize statistics in wooden cubes.

[Text proceeds from here to discuss maps.]

Georg von Mayr (1841-1925) was a Bavarian economist known for his expertise in administrative and bureaucratic statistics. He is know for a novel bivariate map, mosaic plot, and early polar diagrams including a series of star plots.

Read Mayr’s original German text at https://archive.org/details/diegesetzmssi00mayr/page/70/mode/2up

Thanks to Francis Harvey (Leibniz Institute for Regional Geography) for reviewing my translation. Translation ©2025 RJ Andrews. All rights reserved.

About

RJ Andrews helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is currently out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

This summer, we are showcasing a series of historic writing about information design. It is the SUMMER OF CLARITY! These essay inspired the blue marginalia in my new book Info We Trust ($39 from Visionary Press).

Today, enjoy William Playfair’s 1801 preface to his Statistical Breviary. I’ve used the five plates from his book to augment his text. They include his first pie charts. Playfair’s original footnotes are in tiny text.

Preface to The Statistical Breviary

Having about a year ago been requested by the English editor of Mr. Boetticher's Statistical Tables, to consider of some method of bringing them down to that period, without injuring the original work, I proposed to make a supplementary table, comprehending all the countries which have undergone any material change since the publication of the book. I then undertook to make out such a supplementary table; which I did, and it is published at the end of that work.

In the course of executing that design, it occurred to me, that tables are by no means a good form for conveying such information, unless where a number of different countries are intended to be exhibited at once. Where there is only one to be set forth, I can see no kind of advantage in that sort of representation, while the inconvenience of a large size, in a book that is intended to be frequently referred to, is obvious.

I do not conceive that it is in any manner the province of statistical works to contain historical relation, or any thing that is not a simple fact and relative to one single epoch or date. The numbers of people, quantity of ground, revenues, prices of labour, &c. as simple and useful facts, belong to statistics; but the description of the order of the garter, or of the golden fleece, has nothing to do with it. To encumber statistical reports with such information appears to me to be ill placed, and as such improper.

I have composed the following work upon the principle of which I speak; this, however, I never should have thought of doing, had it not occurred to me, that making an appeal to the eye when proportion and magnitude are concerned, is the best and readiest method of conveying a distinct idea.

Statistical knowledge, though in some degree searched after in the most early ages of the world, has not, till within these last fifty years, become a regular object of study. Its utility to all persons connected in any way with public affairs, is evident: and indeed it is no less evident that every one who aspires at the character of a well-informed man should acquire a certain degree of knowledge on a subject so universally important, and so generally canvassed.

Geographical knowledge has long been considered as necessary for persons of both sexes who wish to acquire any tolerable degree of general information; in so much that, next to ignorance of the grammar of one's native language, nothing betrays want of information so soon as ignorance in matters of geography, without which it is almost impossible to carry on conversation long on any general subject.

Geography is, however, only a branch of statistics, a knowledge of which is necessary to the well understanding of the history of nations, as well as their situations relatively to each other. In ancient history, and even down to our own times, there is nothing so imperfect as the accounts given of statistical matters. Ancient historians, and other writers, tell us for example, of great armies raised and great achievements performed; but concerning the finances, and ways and means, they are generally silent. To the importance of this species of knowledge, mankind have only of late years begun to pay a sufficient degree of attention, the want of which, hitherto, leaves us now in great ignorance on many points which it would be very useful for us to know, in order to form a comparison between the ancient state of the world and its present situation.

Statistical accounts are to be referred to as a dictionary by men of riper years, and by young men as a grammar, to teach them the relations and proportions of different statistical subjects, and to imprint them on the mind at a time when the memory is capable of being impressed in a lasting and durable manner, thereby laying the foundation for accurate and valuable knowledge.

Since the value of this study is generally acknowledged, it has become a desirable thing to render it as easy and perfect as may be. In the introduction reasons are given for adopting the mode of representing the magnitude of different countries by proportional circles, but the great tell of its utility is in the mind of the person who takes up the charts. The first of these has been shewn to numbers of persons, all of whom have declared that till they saw it, they had no right and distinct idea of the proportional extent of the different countries such as it gave them. The reason of this is evident: for, as it is not without some pains and labour that the memory is repressed with the proportion between different quantities expressed in words or figures, many persons never take that trouble;— and there is even, to those that do so, a fresh effort of memory necessary each time the question occurs. It is different with a chart, as the eye cannot look on similar forms without involuntarily as it were comparing their magnitudes. So that what in the usual mode was attended with some difficulty, becomes not only easy, but as it were unavoidable.

Whatever presents itself quickly and clearly to the mind, sets it to work, to reason and think; whereas, it often happens, that in learning a number of detached facts, the mind is merely passive, and makes no effort further than an attempt to retain such knowledge.

It would be almost impossible for any person of intelligence to contemplate the first chart without being struck with the great size of Russia and Turkey, and the comparatively small extent of those countries which have borne the principal sway in the world for these last five hundred years, whilst Russia was nearly unknown, and counted but as dust in the political balance of nations. Some general conclusions, accompanied with no small degree of surprise, naturally attend the first view of this proportional chart of nations.

What thinking man who considers the important part that the small republic of Holland has acted, while Russia lay as if congealed in an eternal winter, but will conclude, that if ever the people in those different countries come to be in any degree similar in civilization and intelligence, the importance of the smaller must sink into great inferiority, and in general, that if even the different countries in the world should come to be nearly upon a par in respect of arts, civilization and knowledge, the scale of their importance must be strangely altered, and accordingly it is daily altering: for, as commerce, arts, and civilization, have been making great progress during the last century, the foundation of changes has been solidly hid, and their have begun to take place with unexampled rapidity.

Holland, which was a preponderating power in the beginning, and during a great part of the last century, as it had long before been, entered into the last war shorn of its importance, with the rank of only an auxiliary to France and Spain. It did not long preserve even that diminished rank; for having first submitted to be the tool of a French faction, it was in the course of a few days reduced to obedience by the King of Prussia, who acted with it just as he would have done with a rebellious province of his own dominions; and when the present war broke out, it soon was reduced to what impartial truth obliges us to call a dependant province of France.

Portugal, now so different from what it was in the time when its conquests almost encircled, and did astonish the world, seems to run a risque of sharing the fate of Holland.

Though extent of territory is the ground work of power, as it regulates in a great degree the population of a country; yet we find neither extent nor population will do without revenue: hence we find Poland extensive, populous, rich in soil, and productive, peopled with a race much more zealous of liberty than any of the neighbouring kingdoms, fallen a prey to the power of those very neighbours. The conclusion is, that want of revenue was the cause of its ruin*.

* Perhaps it will be urged that want of unanimity, not want of revenue, ruined Poland; but in answer to this, it may be urged, that want of revenue occasions want of unanimity as well as many other wants.

To render statistical accounts accurate and complete, it is not sufficient that individuals should collect knowledge, and arrange it in order, for the aid of rulers and magistrates. An habitual and regular practice of collecting information, both generally and locally, is necessary; but as vanity is not flattered by employing men to collect such materials, as it does not immediately advance the interests of those who are at the head of affairs, it is to be feared that the business will long be left to the inadequate care of a few individuals.

Where vanity is not gratified, or interest promoted, knowledge is generally neglected. The bushels of rings taken from the fingers of the slain at the battle of Cannae, above two thousand years ago, are recorded: so are the numbers of combatants at the battles of Agincourt and Cressy; but the bushels of corn produced in England at this day, or the number of the inhabitants of the country*, are unknown, at the very time that we are debating that most important question, whether or not there is sufficient subsistence for those who live within the kingdom. We neither know whether the country is increasing or diminishing in population: we are equally ignorant as to its produce, and yet, perhaps, no nation in Europe is better informed on those important subjects than ourselves. No encouragement is given, no proper steps are taken by those who rule, to ascertain points that are so material, while there are Societies instituted for inquiry into matters which are past and gone, rare and useless, or distant and unknown.

* Some efforts have been lately made to ascertain the population of this country, which are entirely inadequate to the purpose, and are therefore to be considered as nothing.

Were the aid and support of public men obtained in collecting statistical knowledge, great progress might be made in it at little expense, and with great facility; but so long as that is not the case, individuals will find themselves reduced to the situation of scanty gleaners, not that of men carrying home an ample harvest.

Statesmen, and those in power, would in the end find themselves amply repaid for any trouble, or moderate degree of experience, that an attention to statistics might occasion; for by that means the operations of government (particularly the revenue department*) would be greatly facilitated. Great statesmen and monarchs have known this in all ages; from whence attempts have arisen to number the people, and take an account of property, &c.

As statistical results never can be made out with minute accuracy, and that, if they were, it would add little to their utility, from the changes that are perpetually taking place; it has been thought proper in this work to omit that customary ostentation of inserting what may be termed fractional parts, in calculating great numbers, as they only confuse the mind and are in themselves an absurdity.

* In the revenue department much accuracy and great attention prevails throughout; but all other national operations are done in a slovenly inaccurate manner, at if revenue alone were worth attending to: it is not so in many countries that are in other respects much worse regulated than this.

Statistical books, like dictionaries, require new editions from time to time, as changes take place among nations; but it is impossible to begin a regular series of such accounts from any period so proper as that just previous to the present war. Europe had been almost stationary for a century, when all at once changes commenced, which, from their nature, their causes, and the general situation of things, will not soon be ended in a solid manner. The first view of European nations is the foundation from which we rise, with an intention to exhibit in a like manner the same nations under the different vicissitudes which the present troubles have occasioned, or in future may occasion.

William Playfair (1759–1823) was a Scottish engineer and political economist. A founder of graphical methods of statistics, Playfair developed several types of diagrams including the line graph, bar chart, pie chart, and bubble plot.

All plates are from the UPenn library. See them in higher resolution, and read the entire Statistical Breviary, here: https://colenda.library.upenn.edu/catalog/81431-p3gh9bs0c

About

RJ Andrews helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is currently out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

This summer, we are showcasing a series of historic writing about information design. It is the SUMMER OF CLARITY! These essay inspired the blue marginalia in my new book Info We Trust ($39 from Visionary Press).

In 1861, Charles-Joseph Minard, eighty years old, published a treatise on his “useful idea”—ingenious flow maps which still inspire today. The essay reveals the design thinking behind his extraordinary intellectual and creative power.

The following original translation continues from part one, which you may reference for more background context. I’ve augmented Minard's original plates with images of works referenced by his text, but not illustrated in the original 1861 publication.

Graphic Tables and Figurative Maps (part 2)

[continued from part 1]

William Playfair’s Tableaux d’arithmétique linéaire, first printed in London and then translated into French around 1787, show colored curves representing annual wheat prices, workers’ wages, etc. [Minard’s facsimile, is Plate IV Fig. 2, above, with a similar Playfair chart from 1821, below].

At first glance, [Playfair’s] wheat prices appear similar to my graphic tables: there are rectangles. But their areas have no statistical significance, while those of my transport route rectangles show the number of tons carried per kilometer. It is such an important metric for traffic that its number is always recorded in the reports.

At the beginning of this century, Layton Cooke published his large tables of colored curves representing the annual prices of wheat, wool, meat, butter, etc. in northern Europe [Cooke example below].

In 1826, I published, using the same system, the various elements of the maintenance of the paving stone of Paris during the last two centuries.

In 1833, still in the same system, Tarbé published tables showing daily deaths in Paris during the cholera of the previous year [Jean Bernard Tarbé de Vauxclairs graphic work on 1832 cholera epidemic was also acknowledged by Marey].

Most engineers recorded daily river water levels in the same way.

These are, among many other applications, those that have been made and are still being made by extending the system indicated by Descartes; but it must be recognized that, whenever the data collected simultaneously have no relationship between them, such as the quantities of goods passing through a canal and the distances they have travelled, the Descartes system has no economic significance nor graphic possibility. Another mode of representing statistical facts was required to compare them by sight alone.

As for my maps, I have heard people say that expressive maps have been made for a long time. However, not only do my maps speak, they count and they calculate for the eye. This is the crucial point; this is the improvement that I have introduced into my figurative maps by the width of the areas, and into my graphic tables by the rectangles.

Other more or less curious representations have appeared in recent years. For example, the number of days with the same prevailing winds in a port was grouped around a compass rose [Minard’s facsimile, Plate IV Fig. 4, above].

In 1827, Baron Ch. Dupin, Senator, and in 1828, Messrs. Balbi and Guerry sought to show the degrees of education and criminality across departments by more or less dark shading [Minard's facsimile, Plate IV Fig. 1, above]; this very ingenious representation was not imitated, probably because it does not give a numerical evaluation. We can say that one shade is darker than another; that is obvious. But to say that it is two or three times as dark is not visible, it is not readable.

As a way of counting by shades, let me mention, and only by analogy, the ingenious process of chemical analysis of water due, I believe, to Gay-Lussac. [French scientist Joseph Louis Gay-Lussac was known for many achievements including his discovery that water is made of two parts hydrogen and one part oxygen.]

In twelve glasses containing the same quantity of distilled water, one dissolves increasing amounts of salt. By pouring in the same quantity of reagent one gets twelve precipitates of increasing intensity. Carrying out the same operation on the water sample to be analyzed gives a precipitate that one can compare to the initial twelve. By conducting the same operation, it is conceivable that two precipitates will have very close to equal color values, which will give the quantity of salt sought. But it is by the equality of color value (which the eye does not mistake), and not by different [salt] intensities, that one obtains a numerical solution.

Let’s return to graphic representations.

Statisticians divided a circle circumference into 365 equal parts, extended the radii going along these divisions. Lengths proportional to daily observations were extended from the circle. Then the ends of these lines were united by a curve, giving a circular representation of all the observations during the year. This is how the weekly revenues of the Western Railway were calculated in 1858 [Minard’s facsimile, Plate IV Fig 3., above].

I saw this mode of representation in the statistical office of the Northern Railway as early as 1846.

In 1851, I used proportional circle areas to represent coal mine production [below]…

… and in 1852, to represent port tonnages [below]. Since this latter representation placed ports with some geographical accuracy, it allowed for comparison by size and position.

The representation by circles leaves much to be desired. By sight alone, it is very difficult to grasp the ratio of the areas of two circles. One must, in a way, go through the evaluation of the diameters whose squares give the ratio of the circles.

But if separate circles are hard to compare, it is easy to compare sectors of the same circle. That is why I used circle sectors to show the tonnage ratios of each port in my map of sea port tonnages in 1857 [below]. One sector represents cabotage [domestic transport, black] and the other foreign trade [blue]. Circle sectors had already been used before I made use of them [William Playfair introduced the pie chart in his Statistical Breviary (1801) and elaborated its use in 1805.].

In February 1847, Alphonse Belpaire, a Belgian engineer, published a notice on his beautiful figurative maps (undated) of the roads, rivers and railways of Belgium, drawn up absolutely in the same system as that of my maps. Belpaire says in this notice that his first map [below] appeared about two years earlier, around February 1845. This is the same this time that I published my first figurative map. Thus, if Belpaire’s assertion is accepted, the two inventions arose at the same time, in Belgium and France.

My publications were imitated more or less immediately. Having sent my 1845 colored graphic table on the Central Canal trade [see Plate III in part 1] to Mr. Comoy, then in charge of that canal, this chief engineer found the idea so good that he continued it, and every year since 1851 he has published graphic tables that imitate mine. He has only added, still following my system, the representation of the size of canal ports.

In a circular dated 20 December 1854, the Civil Engineering Administration recommended the application of Mr. Comoy’s (i.e. my own) graphic tables to the main navigable lines, and this recommendation was followed.

In 1856, Mr. Nicolas, engineer of bridges and roads, compiled interesting graphical tables showing various circumstances under which our railways have operated since their inception. In 1857, Garella, Chief Mining Engineer, published a graphical chart of passenger movement on the western railway network. One of the most successful applications of my figurative map system was made in 1859 by the Chief of Railway Statistics of the Ministry of Public Works; it shows the movement and number of troops sent to the Italian War in 1859. Most recently, the Civil Engineering Administration, adopting my system, published a large 12-sheet figurative map of imperial road traffic in 1856–1857.

In recent years, railway companies have adopted my graphical tables to represent their low-speed traffic. The Orléans company considered a particular aspect of my system, deemed it excellent from the point of view of administration; but it is not up to me to comment on their modifications.

His Excellency the Minister of Agriculture, Trade and Public Works, on the enlightened and benevolent presentation by the Director General of Highways and Railways, was pleased to accept most of my maps. I was able to publish a total of nearly ten thousand copies, graphic tables and maps on various subjects such as passenger traffic on roads and railways; freight traffic in general, and in particular coal, cereals and wine, by water and by rail tonnages of sea ports in France, Europe and world-wide, the consumption of meat in Paris, goods transiting through France and recently the import of raw cotton into Europe, etc., etc.

My figurative maps and graphic tables having been presented to the Emperor [Napoleon III], His Majesty was kind enough, in accepting them, to let me know “that he had viewed them with interest and that he would make use of them.”

Strengthened by this august commendation given to my publications, which are so well received in France and abroad, I feel, at over eighty years old, the satisfaction of having given birth in my old age to a useful idea, of having propagated it through many applications, and of thus still doing some service at the end of my career.᠅

Charles Joseph Minard (1781–1870) was a French civil engineer and professor known for his significant contribution in the development of numerical data on geographic maps, especially flow maps.

This translation was first published at InfoWeTrust.com. I am grateful to William Berg, Professor Emeritus as Wisconsin-Madison, for improving this translation effort.

Translation ©2019, 2025 RJ Andrews. All rights reserved.

Let me know if you can help improve this translation, or have suggested entries for future editions of THE SUMMER OF CLARITY.

About

RJ Andrews helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is currently out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.

This summer, we are showcasing a series of historic writing about information design. It is the SUMMER OF CLARITY! These essay inspired the blue marginalia in my new book Info We Trust ($39 from Visionary Press).

In 1861, Charles-Joseph Minard, eighty years old, published a treatise on his “useful idea”—ingenious flow maps which still inspire today. The essay reveals the design thinking behind his extraordinary intellectual and creative power.

Minard was eighty years old when he wrote this treatise. He probably thought he was at the end. We now know that Minard would go on to produce two dozen more flow maps before his death a decade later. They include a beautiful series on cotton imports during the American Civil War, a map showing the spread of ancient languages, and an iconic depiction of Napoleon’s disastrous Russian campaign.

The following original translation augments Minard's original plates with images of works referenced by his text, but not illustrated in the original 1861 publication. Further context is given by me in [square brackets].

Graphic Tables and Figurative Maps (part 1)

“To each his own”

Today’s great expansion of statistical investigation has made it necessary to record its results in less dry and more useful forms; forms that can be explored faster than numbers alone. Hence the various representations that have been imagined, including my graphic tables and figurative maps.

By giving statistics a visual direction, I followed the general impulse of the mind toward graphic representations.

Today, images apply to everything: Real estate, furniture, and clothing advertisements are always accompanied by plans and drawings. In major criminal trials, the layout of the premises is put before the jurors. The holiday gift book is more the work of the lithographer than of the writer. La Fontaine made animals speak in his fables; but we wish, even more, for Granville to show their likeness.

In the expectation of publication, a photographer is attached to every expedition; it’s not just science books, as was sometimes seen even two hundred years ago, that show text and image together on the same page.

Illustration invades everything, and by making statistics visual, I satisfied the need of the day. But did I sacrifice to the taste of the time? Or, have I contributed to increasing the usefulness and shortening the time of interpreting statistical studies? That’s what can be judged by reading the following pages.

The dominant principle that characterizes my graphic tables and figurative maps is to make numerical proportions immediately appreciable to the eye, as much as possible.

I first imagined, in 1844, a graphic table of which Plate I is a copy [below]. Published in one hundred copies, it represents the number of passengers on the total and partial routes of several railways.

On the horizontal line I added, end to end, the lengths between stations. These lengths are the bases of as many rectangles whose vertical heights are proportional to the numbers of all passengers who passed between two consecutive stations.

A portion of these rectangles corresponding to the passengers traveling the total route distance is distinguished by a darker shade than that of the portion relating to the passengers who traveled only part of the total route.

By means of these rectangles and shading, one can see at a glance the ratio of all travelers who passed between any two stations and the ratio between total route and partial route passengers.

Additionally, the total area of each rectangle is proportional to the number of passengers carried 1 kilometer (an important metric of railway traffic). The comparison of these areas promptly give, in one image, what written numbers only give slowly by arithmetic multiplication.

The following year, in 1845, I published, using the same system, the first graphic table on the trade of the Central Canal [Canal du Centre was the first canal allowing boats to pass from the north of France to the south]. Plate III is a facsimile [below]. In this graphic table the rectangles relating to various goods have different colors. This lithographed and colored graphic table was distributed at the end of 1845 to about a hundred people: engineers, deputies, etc.

This mode of representation has a small disadvantage in terms of commercial use: sometimes the progress of a commodity cannot be easily followed by the eye, because the rectangles of the same color that represent it are separated by those of another color. If we want to be fully aware of the movement of this commodity, we must have a graphic table for it alone.

In that year, applying the same system differently, I arrived at the figurative maps in which the rectangles of the graphic tables are replaced by tinted areas that follow the transport routes and whose widths are proportional to traffic.

It March 1845, during the [planning] discussion of the railway from Dijon to Mulhouse, I published my first figurative map; Plate II is a facsimile [below]. Passengers traveling on the roads in the region, between two cities, were represented by tinted areas that follow the contours of the roads and have widths proportional to the number of passengers passing through them in public coaches. Mr. Frémy, then an auditor with the State Council and now a State Councillor and Governor of Real Estate Credit, carefully took note of the situation on the spot.

My lithographed map, in two hundred copies, was distributed to the members of the Council of Bridges and Roads, as well as to Representatives of the Departments concerned, to Engineers, etc. and enlightened the meeting’s discussion [Departments are government jurisdictions below administrative regions]. The [figurative map] idea was broadly welcomed to guide opinion in the choice of a route in general, and in particular the route which followed the Doubs valley [see how more coach traffic flows along the Doubs River compared to the Ognon River]. The map was so successful that someone produced, using my name, a forgery where he drew false zones in favor of a rival line through the Ognon valley.

It is by sight alone that this map, which was found to be eloquent, made visible the relationship between the numbers of travelers because, you may notice, it does not include a single numeral.

Advantageously, an instant evaluation of the reported statistical results and their overall configuration can be judged at a glance with figurative maps, as long as they are contained in the ordinary field of vision, while maintaining respective geographic positions.

These benefits were readily grasped by the engineers. A few months after my figurative map of passengers between Dijon and Mulhouse had appeared [Plate II, above], several engineers imitated it to support their railway projects: Mr Reynard for the railway from Grenoble to Valence, Mr Léon for the railway from Poitiers to La Rochelle, Mr Maillebiau for the railway from Bordeaux to Bayonne. All produced, based on my system, figurative maps of the travelers circulating in these regions.

Nevertheless, if my maps immediately afford simple ratios such as double, triple, etc.., which, I repeat, is only obtained from numerals through a rather long mental operation, this advantage fades if you have to compare very disproportionate elements. With numerals, when the smallest has only two digits, with a little time the mind can grasp the quotient; which is why I also write the numerals on several of my maps.

Additionally, the spontaneous evaluation of large quantities has always been very uncertain. Who doesn’t know how difficult it is for a general to count enemy troops that are in sight? Napoléon himself stated that even with his long experience, he was mistaken by a quarter at least. And yet nothing can tell the eye the number of men better than the men themselves placed next to each other.

In the past, I have not seen any expressive maps other than those of the first geographers who depicted countries by their animals.

The application of algebra to the geometry of Descartes gives us geometric curves representing the relationships between two variables. Tidal curves were one of the most remarkable applications of this discovery; I do not know if we know to whom they are due. This is an improvement in graphical representation, since the eye can see the successive heights of the tides.

Ons-en-Bray’s ingenious 1734 idea to draw a graphic table of prevailing winds using the wind itself acting on a mechanism can be applied to the representation of the tides. [top of Ons-en-Bray’s anemograph shown below]

Today we have the tide gauge, an instrument by means of which the sea itself draws on paper the heights of the tides across successive hours, giving the most complete and accurate curves of this movement. [The first self-recording tide gauge was created by James Mitchell and installed in 1831 at the Royal Navy’s Sheerness Dockyard in Kent, England]

In the 18th century, we see the first application, I believe, of Descartes’ system to phenomena without determined laws, such as those concerning public economics. Descartes drew curves according to the relationship of two variables. The question was reversed to investigate whether simultaneous, more or less correlative facts, recorded on abscissa and ordinate, could correspond to known curves or give useful representations. [The abscissa and the ordinate are respectively the first and second coordinates of a point.]

Minard’s treatise is continued in part 2:

Charles Joseph Minard (1781–1870) was a French civil engineer and professor known for his significant contribution in the development of numerical data on geographic maps, especially flow maps.

This translation was first published at InfoWeTrust.com. I am grateful to William Berg, Professor Emeritus as Wisconsin-Madison, for improving this translation effort.

Translation ©2019, 2025 RJ Andrews. All rights reserved.

Let me know if you can help improve this translation, or have suggested entries for future editions of THE SUMMER OF CLARITY.

About

RJ Andrews helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s book, Info We Trust, is currently out now! He also published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.