by R J Andrews
Tables are a foundational data visual. They originated as multipurpose tools to store and communicate numbers. Old paper tables, like John Graunt's Bills of Mortality, survived so future strangers could reference the data. We no longer need to rely on visual tables to store data. Today, the only goal of the published table is to inform the reader, not to store precise values for future calculation
Basic problem: Make data more easily and effectively handleable by minds.
JOHN TUKEY, 1977
Graunt invited London to discover how useful a data table can be. The world's interest has not abated since. His tables made data more easy to handle by using only simple, printing-press type: letters, numbers, lines, and spacers. The same elements continue to be the building blocks of much data reading today.
The storage location of the original data is revealed by giving each value its own cell. Good table design augments our ability to read data. It also acknowledges the natural limits of human comprehension. For example, aligning value decimal points allows comparisons to be made quickly with a vertical scan.
A table is for communication, not data storage … for human eyes and human minds.
HOWARD WAINER, 2007
Be careful of how many numerals express each value. As Howard Wainer advises, “Round—a lot!” Two numerals is a good target. Reduce decimal places. Switch units to thousands or millions. Do what it takes with numerals to make numbers more intelligible.
Further: More precision than two numerals often cannot be statistically justified.
Aggressive rounding may reduce precision by a startling amount. But consider that we cannot actually appreciate more digits easily. What does the precise difference between 10,564.4 and 82,935.34 actually mean? If we compare these two numbers in our mind we retain only the gist of each, not the detail. The mind might understand that the larger number is about eight times bigger than the smaller. Skip the illusion of being able to comprehend more than you can. Express them in thousands: 11 and 83.
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.
WILLIAM PLAYFAIR, 1801
Superfluous precision clouds our ability to make sense of a wall of numerals. The visual weight of a seven-numeral string, like 82,935.34, reduces the impact of the numeral that matters most, the leading 8. A sheet full of similarly precise seven-numeral numbers would be impenetrable.
Thoughtful rounding and spacing improves tables. Horizontal and vertical lines can help too. Lines can communicate meaning. Do not line every cell; instead reserve them for emphasis. John Tukey advised that double lines between things should add up, like one big equals sign. Single lines should separate summary values from their input data. Heavier lines may highlight cells of special importance. These lessons applied to our slice of the population data yield an improved table, now in thousands:
Boxes, bolding, and shading can be used to direct attention to values of particular interest. Table color meaning varies across cultures. In accounting, black is positive and red is negative. Spreadsheet colors may indicate what type of value each number is: output of a formula, a hard-coded constant, an alert, or a link to another sheet.
The equals sign was first described by Robert Recorde in 1557: “to avoid the tedious repetition of these words: “is equal to”, I will set (as I do often in work use) a pair of parallels, or Gemowe lines, of one length (thus =), because no two things can be more equal.” Gemowe means twin, from the same root word as the constellation Gemini.
Table entries are often arranged in alphabetical order. This default is an outdated way of presenting information, developed by scholars to catalog work for easier retrieval. Today, alphabetic order lingers in a technological age that has made search easy. Instead of the first letters of some label, try to sort categories based on their values. It can reveal groups, voids, and other simple patterns.
Toussaint Loua introduced the shaded table in 1873 to ease exploration of Parisian demographic data. His résumé graphique shaded cells darker as their value increased.
Spaces between rows can emphasize value clusters. Some of these practices are applied to the full U.S. Census population data for Columbia County to show how rounding, formatting, sorting (by 2010 population), grouping, and using lines to separate summary values can shape data into better information.
1,000 is represented by K for kilo, as in kilogram, a derivative of the Greek word for thousand (Χíλɩοɩ) created in 1795 France.
Can tables be pushed to convey even more? Recall that the origin of numerals lies with the tallying of observations. What if we tally each observation according to its value from the get-go? John Tukey adopted this strategy with his stem-and-leaf display. It tallies low-precision values in a way that invites group profiling. The same town populations from above are tallied below with a stem-and-leaf. For example, the Town of Kinderhook's 2010 population of 8,498 is rounded to 85**. To record Kinderhook on the chart, work your way up from the bottom of the thousands (k) column, the stem, until you come to 8, which corresponds to Kinderhook's thousands digit. Then, staying in the 8k row, place a 5 (Kinderhook's hundreds digit) in the hundreds column, a leaf. Beside the leaf, place the town's name. The stem-and-leaf display is just one example of the creative possibilities of using thoughtful design to better present information.
Perhaps we still see so many tables because they match how we think about data's storage structure. Why not just print out the map of where all the data live? I think readers place their trust in tables because tables seem to give direct access to the raw data.
In the course of executing that design, it occurred to me, that tables are by no means a good form for conveying such information … I can see no advantage in that sort of representation.…
WILLIAM PLAYFAIR, 1801
Indeed, tables persist as tools for looking up values in small datasets. Tables also aid simple spreadsheet arithmetic and continue to be popular visual accessories. A tiny table in the corner of a thematic map seems to prop up a general confidence by saying, Look there. See? We used numbers to make this. Tables appear to be a transparent way of presenting the data. No values are hidden, but unfortunately, not much else can be seen. Looking at tables of any substantial size is a little like looking at the grooves of a record with a magnifying glass. You can see the data but you will not hear the music.
CHAPTER
3
EMBODIED
What our eyes behold may well be the text of life but one's meditations on the text and the disclosures of these meditations are no less a part of the structure of reality.
WALLACE STEVENS, 1951
To create better information, we must enter into a dialog with data. Creating information and exploring data are partner endeavors. Together, these activities form the humanize-probe cycle that powers all data storytelling. To wade into this cycle is to submerge yourself into the pool of data chaos, resurface to consider what you found, and dive in again with a better idea of what treasure might lie beneath. First, let us get our toes wet by pausing to consider how we think about, and navigate about, the world of our experience.
The hero has to go … “where the night and the day meet together”
MIRCEA ELIADE, 1952
Stretch Cognition
When we encounter a never-before-seen object or meet a strange situation, we mostly do not panic. Instead, we stretch what we already know to help us understand what is going on. Our ability to extrapolate what we know to what we don't know is the secret of our success. This power—called analogy-making, category extension, and metaphor-mapping—helps make strange things familiar.
The ceaseless activity of making mappings between freshly minted menta
l structures (new percepts) and older mental structures (old concepts)—the activity of pinpointing highly relevant concepts in novel situations—constitutes the analogical fabric of thought, and the unceasing flurry of analogies that we come up with is a mirror of our intelligence. … the human mind is forever driven to transform its categories, not just to use them as givens, [these] intellectual advances are dependent on conceptual extensions.
HOFSTADTER AND SANDER, 2013
All of your mental categories, or concepts, are related through a dense network of relationships between your brain's neurons. In a novel situation, this network comes to your aid. It surfaces patterns from your long-term memory that might best help figure out what is going on. The first beanbag chair you ever saw was an ugly lump on the floor, unlike any furniture you had seen before. But once you saw that the category chair could be extended to include this new lumpy object, all kinds of chair-properties, namely that chairs are good for sitting, were suddenly mapped to the giant beanbag. The moment one realizes chair can be stretched to the lump, the category of chair also expands. Chairs now include lumpy beanbags. Every new encounter offers the possibility of extension.
Our brain's network helps us understand strange situations. It also filters the avalanche of stimulation that the world barrages us with, to focus us on what is important. The world presents too many objects at once for us to take in. Mental categories help make sense of the cacophony of the world by directing attention to what is useful, now. Knowing what is important allows us to ignore the rest. Many things evade our notice, by design. The alternative would be to suffer like Buridan's donkey: In a constant panic of information overload, never sure of what to do, we would be unable to take any action.
Buridan's ass is the parable of the thirsty and hungry donkey who, unable to choose between the equally appealing bucket of water or pile of hay, dies from the parched starvation of analysis paralysis.
Our experience of the world is naturally biased. We tune our senses to what our existing concepts consider important. A practical prejudice, which focuses our attention on what we consider useful, is necessary if we are to swim through the chaos. Along your path, interesting things attract your attention. The most compelling and novel stimulation helps cultivate and grow your personal microcosm. Some learn the hard way to not touch a glowing stove-top, but their updated model of how the world works prevents them from doing it again. The coherent reality you inhabit gets refined. By paying attention to some things, and ignoring the rest, we each thread our individual life through the world of experience. Each of our paths is a little unique slice of reality, a microcosm, of the common world we all share. It is our nature to strive for ever-more useful connections to help us better interact with the world.
“You don't know much,” said the Duchess; “and that's a fact.”
ALICE'S ADVENTURES IN WONDERLAND, LEWIS CARROLL, 1865
Curiosity is the search for more effective categories. It often helps us also achieve a more accurate understanding too. But everyone's cultivated personal microcosm is first about guiding their own life. One quick way to gain a more effective understanding is to compare your own unique microcosm to others. No two people's lived realities are quite exactly the same. Seeing how others perceive the greater world is a route to your own improvement.
All that has been learned empirically about evolution in general and mental process in particular suggests that the brain is a machine assembled not to understand itself, but to survive. … It throws a spotlight on those portions of the world it must know in order to live to the next day, and surrenders the rest to darkness.
E.O. WILSON, 1998
Evaluating yourself against others is productive when it helps you identify holes in your personal experience of reality. Others can teach us how to refine our microcosms so that they are more true, and more effective. From this vantage, we can imagine our way to empathy and compassion. Together, we can close the gaps between our microcosms and inspire one another to stretch our minds in new, more useful directions. Ideally, we strive to understand the microcosms of others, and be generous and kind with how we share our own perceptions of reality. A parent not only safeguards the kitchen, but also teaches their child about its dangers, such as sharp knives, electrical outlets, and hot stoves.
From the experientialist perspective, metaphor is about imaginative rationality. … New metaphors are capable of creating new understandings and, therefore, new realities.
LAKOFF AND JOHNSON, 1980
Imagination lets us exceed our inevitably limited point of view to find perspectives not in existence or dimensions not yet accessible. … Reaching across the gap to experience another's way of knowing takes a leap of imagination.
NICK SOUSANIS, 2015
So, what does this all have to do with data? A data story is a new microcosm for the reader to consider, inhabit, and compare against their current worldview. It is one way to generously share a perspective on how the world works. Whether it is a stark bar chart or a fully interactive experience, each data story invites the reader to take a new look. It is a real privilege to build and offer visual slices of the world for a reader's consideration. We hope that each data story can help others better understand what is actually going on. We hope to nudge one another toward better personal perspectives and, in the process, maybe even change who we are.
The information gap theory of curiosity, first advanced by George Loewenstein in 1994, interprets curiosity as “a form of cognitively induced deprivation that arises from the perception of a gap in knowledge or understanding.“
If data stories compel readers, it is because their territory is somehow interesting. People crave knowledge. We are interested by things that help us achieve more useful categories, which, in turn, help us better interact with the world. If readers sense that a data story can help them improve their mental map of how the world works, then they will open themselves up to being informed. As data storytellers, our job is to help readers in their quest to enrich their categories. We do this when we provide interesting information.
Interesting theories are those which deny certain assumptions of their audience, while non-interesting theories are those which affirm certain assumptions of their audience.
MURRAY DAVIS, 1971
Only certain information can actually extend someone's knowledge. If it is too familiar the reader will be bored; too strange, and the reader will disengage. We ignore what we have already accounted for and what we cannot possibly understand. In 1971, Murray Davis characterized interesting things as those that credibly challenge what is taken for granted. His simple script promises your information will engage with an audience's existing mental categories: You probably think this, but it's actually not true. Here's the proof, and here's why it matters to you.
Stories that simply challenge the status quo certainly can attract a lot of interest. But this is only one way to make something interesting. Above all, we must appreciate that to engage an audience is to activate the frontiers of their existing mental categories.
For we, on our little pile of mud, can only conceive of that which we are accustomed.
VOLTAIRE, 1752
At their best, data stories address readers where they already stand, and expand their worldview to someplace new. Making things simple is not how this is done. The mind is phenomenally complex. It deserves more respect than just dumbing down the data. Instead, we must aspire to meet readers where they already perceive the world. This way, we can inform readers in a way that lets them integrate new knowledge with their existing concepts.
We are at our human best as creatures of the shore, with one foot on the hard ground of fact and one foot in the sea of mystery. … It is at the shore that the creative work of the mind is done—the work of the artist, poet, philosopher, and scientist.
CHET RAYMO, 1998
Consider what concepts readers arrive with. What categories, analogies, and metaphors already steer the way they perceive? What do they
anticipate? How can we account for their prior experience as beings with an embodied cognition who have lived the world?
Embodiment expresses an abstract idea concretely, especially via relation to a person's body.
Good data stories reflect real life. But not just because we hijack a superficial set of sensory organs used by our ancestors to look across the savanna. Good data stories work when we meet minds right where they already are, already surfing invisible worlds.
Metaphor-Mapping
Today, we are wanderers of invisible worlds. Each of us is centered by a personal identity, which cannot be seen. Its daily experience is a struggle with a swirling tangle of forces we cannot see. We have all yearned to advance our self's economic and social worth, if only we could grasp money and love.
Analogy is the fuel and fire of thinking.
HOFSTADTER AND SANDER, 2013
New metaphors emerge when two ideas fire at the same time and their resemblance becomes conflated. Metaphors help expand understanding by linking the unknown to the known. As we map the unfamiliar, our conceptual sphere, our microcosm, expands. Lakoff and Johnson introduced conceptual extension with the metaphor that ARGUMENT IS WAR: Your claims are indefensible. He attacked every weak point in my argument. His criticisms were right on target. I demolished his argument.