Monday, July 22, 2019

Absolute or relative values?

In visualization sometimes the simplest choices are the hardest ones to make. My favorite example is whether to show absolute or relative figures. Take this map by the Urban Institute. It displays what would happen if the Affordable Care Act were repealed. It's the graphic chosen to promote this report on the Web and in social media.

Montana (+177%), West Virginia (+176%), and Maine (+165%), would witness the largest increases, but that's because their populations are small (1.0, 1.8, and 1.3 million). Not surprisingly, the largest absolute increases would happen in more populous states: California (+3.8 million insured), Texas (+1.7 million), and Florida (+1.6 million). You can see the data here. These are the top 10 states in absolute terms:

What's the right choice, total counts or relative values? This is always a decision I struggle with, and my answer is often both. On one hand, it's advisable to use adjusted data—percentages, rates—when designing a choropleth map, but is that map alone enough? Why are 83,000 Mainers and 112.000 Montanans represented by a darker color than 3.8 million Californians or 1.6 million Floridians?

Moreover, what's more informative to someone interested in a topic like this, the relative change or the total number of people who would be left without health insurance if the ACA disappeared? I'd choose the latter, but we should never assume that our preferences are representative of a majority of viewers.

Wednesday, July 17, 2019

Keep those legends

There was a debate today around maps from a Washington Post investigation about the opioid epidemic.

Some people praised them because they lack legends (other maps in the same story do have them). They argued that legends aren't sometimes that necessary because most viewers don't look at them anyway, as they just focus on the overall more-less message of the chart, not on its details. An article used in the debate even suggests removing sources.

I'm writing this on my phone, as I'm not in front of my computer today, so allow me to be brief: please never remove sources, and consider keeping your legends, particularly if you can make them unobtrusive (small, or placed behind an on/off button) and simple. Think twice before getting rid of them.

Legends aren't optional add-ons. They are an integral part of visualizations. I agree that a legend in a graphic aimed at a general audience shouldn't be overly detailed, but removing it completely may be going too far in most cases (I'll concede, though, that it might be acceptable in certain specific situations, including the map in the discussion, or these).

It turns out that I'm reviewing the results of some controlled observations I was part of, and I've seen that (a) many viewers can and do ignore legends but, (b) some others refer to them when reading a visualization, and feel frustrated and distrust the graphic if they can't find what they need. Contrary to what some of you may believe, this isn't dependent on education: people in this latter group didn't necessarily have college degrees. If you can place a legend on the side or at the bottom of your graphic, why wouldn't you serve them? Let's be cautious.

Saturday, July 13, 2019

Nonsensical diagrams

I have a soft spot for nonsensical diagrams. I don't cover them in-depth in How Charts Lie, but I have a small collection in my computer that I return to when I'm in need of a good laugh. This one is from Sebastian Gorka's PhD dissertation:

Here's my favorite from the Gorka subfolder; I'm no expert in geopolitics, but I'm quite certain that the “mechanics of terrorism” are more complicated than this:

Some of Gorka's visualizations are puzzlingly minimalistic, making me wonder whether they are necessary at all. See this beauty:

Gorka writes about this one that it's “frighteningly complex” and that it defies “many conventional wisdoms.” It does indeed!

This is the structure of Al Qaeda; I wonder whether a PhD dissertation shouldn't be a tiny bit more specific:

It's easy to make fun of Gorka, who is just a boor and a grifter, but nonsensical diagrams also appear in best-selling books. In 2016 Rolling Stone's Matt Taibbi launched a contest to ‘Make the Most Meaningless Thomas Friedman Graph’, which has its own Twitter hashtag, #FriedmanGraphs (don't miss it). Taibbi was spoofing Friedman's book Thank You for Being Late, which showcases graphics like this:

My favorite from Taibbi's contest isn't the winner, but this one:

Self-help guru Jordan Peterson is famous for his YouTube lectures and his doorstop 12 Rules for Life. I suspect many of his fans haven't read his previous Maps of Meaning. The diagrams in it were described on my Twitter feed as Dungeons&Dragons campaign maps:

(FYI: I try to design far better maps when I play Dungeons&Dragons).

Nathan J. Robinson called Peterson's graphics “masterpieces of unprovable gibberish”. He has a point:

Some of Peterson's diagrams are pretty heavy metal:

Footnote: A while ago I made fun of Peterson's diagrams and one of his fans replied that it was unfair to critique them without reading them in context. I agree, but it turns out that I did read Maps of Meaning, and can understand its philosophical references. This doesn't make the diagrams better, but even funnier.

Thursday, July 11, 2019

Updated website and calendar of talks

My collaborator and student Yuan Fang has been quietly redesigning my website, It's now more focused on the books and consulting, and its style is simpler. We've added a calendar of talks that I'll update regularly; I also have a Google sheet with the same information.

You'll see many events related to How Charts Lie close to its publication date, October 15. I'll be visiting New York, DC, Boston, and many other cities (there'll be book signings at some of these places):

Also, How Charts Lie is just three months away, so my publisher, W.W. Norton, is already promoting it. They've created a humorous flyer that reflects the book's lighthearted tone, and that has already appeared on its Amazon page.

Tuesday, July 9, 2019

A mosaic plot that exemplifies good design practices

Bloomberg's Lauren Leatherby and Chris Martin visualize energy consumption per country, and their reliance on coal, natural gas, petroleum, nuclear, and renewable sources. The result is an interactive mosaic plot (also known as “Marimekko chart”) that reveals plenty of insights, such as the rise of China and India, or the meager increase of renewables since 1980.

This visualization exemplifies a few interactive design good practices. For instance, if you search for a country, a black outline appears in front of it. This outline remains visible while you play with the time slider, allowing you to focus on that country. Here's Japan:

Moreover, when you hover over any country, a time-series stacked area graph appears, so this visualization also applies a principle described in The Truthful Art, and that I also explain to the general public in How Charts Lie: when a topic is complex, we can't expect that a single visualization will suffice; instead, we may need to encode the same data in multiple ways, as each of them may yield different patterns or trends.

Bloomberg's mosaic plot is part of a long story that showcases other charts, and that discusses how energy production and consumption have varied in several countries in the past few decades. Don't miss it.

Friday, July 5, 2019

Visualizing the unfamiliar

BBC World Service's Josh Rayman has published an article about their efforts to visualize the parliamentary elections in Afghanistan. The piece explains how he and his team overcame the challenge of being unfamiliar with the story they wanted to cover.

This is a very common problem, so the article is educational: it demonstrates how thorough and careful we need to be when facing a situation like that. Josh explains:
We started data analysis a long time before approaching our language teams about working on the project, looking for interesting data stories and then reaching out for the context. The initial data set was tens of thousands of lines: one row for each candidate voted on in each ballot box, with a great number of empty rows. Over half the candidates appeared on at least 100 ballot locations, with some candidates in Kabul standing in over 400. It was daunting to know where to start. 
The raw data didn’t indicate winners, so we created a script that calculated the winning candidates, and accounted for the women’s quota. We also had candidate gender stats manually calculated by our producer from BBC Pashto who went through all 2,500 candidates to create region-by-region numbers. 
We aggregated the data using Python (plus pandas) and node.js to make more manageable JSON files for the dashboard. We were able to slot in new data as it arrived over a long period of time, and we could combine the partial 2018 dataset with the existing 2010 data on the fly.

I think that it was NYT's Amanda Cox who suggested in a talk that the best visualization stories often lurk in data that is hard to obtain, that you need to generate yourself, or that is publicly available, but also difficult to understand. Josh's article is yet another example of that.

Thursday, July 4, 2019

Mapping diversity and taking probability and base rates into account

It's 4th of July, so let's celebrate diversity instead of tanksLazaro Gamio has written about his latest interactive visualization, a nice map of the United States showing where diversity has increased and decreased:
America is becoming more racially and ethnically diverse, on its way to becoming majority non-white in 2045 — but some parts of the country are changing more rapidly than the rest [...] Using Census data, we calculated a diversity index for every county in the United States going back to 2009. Each number represents the probability that two people chosen at random will be of a different race or ethnicity.
At first I was surprised at first that Miami-Dade county's diversity decreased a bit in the past decade, but then I realized it makes sense:
Miami-Dade County (49.9) is less diverse than the country as a whole, mainly because 70 percent of its population is Hispanic.
And what about this?
The counties seeing the greatest relative increase in racial and ethnic diversity are among the least diverse places in the country — particularly in the Midwest.
This may be a matter of probability. If a county's diversity is already very low, it's more likely that it'll increase a bit than it is that it'll decrease even further. The same may be true for the reverse situation: if a county's diversity is very high, it's more likely than not that it'll decrease in the future.

I also got curious about how much the index varies depending on population size. It may be that the largest and smallest changes happened in sparsely populated counties. If you have a county with 1,000 inhabitants and 100 of them are minorities, the probability of choosing a minority person at random is 10%. If 50 new minority people move in, the probability increases to 14% (150/1,050 = 0.14). But if those same 50 people moved to a county with 10,000 inhabitants, their impact on diversity wouldn't be that large.

The data Lazaro used is available, so you can play with it. Here are the results, excluding the three counties that had increases of more than 1,000%, Tucker, WV, Owsley, KY, and Hand, SD (see interactive version here):

Wednesday, July 3, 2019

Let's design more explanation graphics —but let's edit them more tightly

El Mundo's Emilio Amade has announced their latest project, an impressive scrollytelling infographic and narrated animation (below) commemorating the 50th anniversary of the Apollo 11 mission.

Years ago, my colleague Hiram Henríquez published his Master's thesis, “The Importance of Explanatory Infographics” (PDF), in which he chronicled the steady vanishing of pictorial 2D, 3D, and animated explanations from news media, which parallels a substantial increase in the use of data visualization. I discussed some of the reasons in Nerd Journalism.

Pictorial explanation infographics, both static and animated, are still around, but their presence is greatly diminished in the news because they are expensive and time-consuming to produce —and also sometimes to consume: El Mundo's Apollo 11 project is an example; it's marvelous, but I wonder whether it goes into too much detail and takes a bit too long to read, a trap I repeatedly fall myself into when writing articles and books or designing visualizations. I keep reminding myself that graphics are powerful when they are not only beautiful, but also tightly edited. In any case, I hope this type of project will eventually make a big comeback everywhere, as I enjoy explanation graphics a lot. Just take a look at this:

Tuesday, July 2, 2019

Visualizing unreported murders in Mexico

El Universal, in collaboration with Google News Lab, has just launched a project that analyzes unreported murders in Mexico, those cases that don't get covered in the news. El Universal's deputy managing editor, Esteban Roman, explains in a making-of article:
Our first step was to establish a process to determine the absence of news. We explored articles on violence to understand how they compare to the government's official registry of homicides. In theory, each murder that occurs ought to correspond with at least one local report about the event. If we saw a divergence, or if the government's reports were suddenly very different from local news coverage, we could deduce that journalists were being silenced. 
Early on, sorting through news articles seemed impossible. We knew we needed to find a news archive with the largest number of publications in Mexico possible so we could track daily coverage across the country. Google News’ vast collection of local and national news stories across Mexico was a good fit 
The effort required us to identify the difference between the number of homicides officially recorded and the news stories of those killings on Google News. This required machine learning algorithms that were able to identify the first reported story and then pinpoint where the event took place. With that information, we were able to connect reported events by media with the government's reports on homicides across more than 2400 municipalities in Mexico.

El Universal's algorithm detected several “silent zones” or “news deserts” (their terms), regions where there's a great disparity between the murders that appear in official statistics and those reported in news media. The resulting story, maps, and graphs show the total number of murders per year in comparison to rates and news coverage.

Monday, July 1, 2019

Aesthetics and ethics: Jaime Serra's visualization philosophy

Jaime Serra is the best visual information/data artist I've ever met, and the only reason he isn't widely known outside the Spanish and Portuguese-speaking worlds, I think, is that he doesn't speak English and he hasn't translated much of his work.

This post is just an excuse to (a) make you aware of his career, which spans over three decades, and (b) to promote an online course Jaime has recently launched where he explains his thinking and methodology.

The course is interesting and affordable ($20). It's in Spanish, but videos are subtitled in English and Portuguese, and judicious use of a translation program such as Google Translate will help with the website itself.

In the lectures Jaime discusses dozens of visualizations and talks about his influences—spoiler: most aren't designers or visual artists, but musicians, poets, and psychologists—and his philosophy. For instance, he argues that in the kind of visualization he favors, aesthetics and ethics are inseparable in the sense of making form and appearance anticipate, reinforce, and match the themes and content of any graphic. One of Jaime's obsessions is to avoid templates, boredom, and repetitiveness; each visualization should be unique, unapologetically subjective, and have its own personality. Giorgia Lupi has said something similar.

Play the intro video of the course to see the breadth and depth of Jaime's work:

Friday, June 28, 2019

Lines aren't just for time-series

The other day there was an interesting debate on Twitter. The Star Tribune's C.J. Sinner posted the graph below and wrote: “Repeat after me: Line charts should be a time series! Line charts should be a time series!  Found in @cityofsaintpaul (MN) 2040 Comprehensive Plan doc.” This is the source; the graph is on page 13. See it below.

The line represents the percentage of people in each district who are people of color (POC), and the bars correspond to the percentage of participants in several community engagement events held in those same districts who were also POC. Sometimes these percentages are close to each other, and sometimes they aren't:

replied to C.J. for two reasons. First, because even if I don't think that this graph is great—more about that later,—I don't find it misleading. Second, because I don't think that “line charts” should always be time series.

Let's begin with the latter: it all depends on what we mean by “line chart”. That term is often used as a synonym for time-series line graphs, which indeed should display time on one axis. That's the strict and traditional definition of “line chart”.

But we shouldn't infer from that definition that all graphs that use lines varying in length or height should also display time on one axis. Think of density curvesconnected scatter plots, or parallel coordinate plots. In the new book, How Charts Lie, I explain that I've met people who find these types of visualization confusing because they apply to them the wrong mental model—that of a time-series line chart—and they feel frustrated.

The impulse of some designers is to say “let's not frustrate any reader; let's use lines just for time-series.” That's self-defeating. If you think that a novel or unusual graphic form is the best way to tell a story, but that some readers won't understand it, don't refrain from using it; instead, explain how to read it.

As for the graphic that illustrates C.J.'s tweet, I confess I was playing devil's advocate in the discussion. I agree that going against expectations and conventions when designing a visualization is often risky, and that it's not justified in this case—but it is justified sometimes.

However, my problem with the original chart isn't, as some argued, that the line might make readers infer some sort of spurious continuity between districts. It seems to me that once you read the graph's legend and labels—something we must always do anyway—it's obvious that we should estimate percentages based on the vertical position of the points connecting the segments of the line, not on the slope of those segments.

My problem is that the line doesn't seem to be necessary, and it's not very efficient at letting you estimate percentages and differences. We could redesign the chart as a bullet graph:

However, let's suppose that we sort the districts from highest to lowest percentage of people of color, and then we add a second variable, such as percentage of people who are 65+years-old (warning: the numbers aren't real; I made them up.) Imagine, just for the sake of argument, that one of the purposes of the graph were to show that the lower the percentage of POC is in a district, the higher the proportion of elderly people becomes:

In a case like this I wouldn't oppose adding connecting lines to emphasize this higher-lower/lower-higher pattern, even if it's unorthodox. Not because the lines encode anything (they don't) but because they can be perceptual aids that highlight one of the key messages of the graph:

Going back to the original graph, here's another alternative design: if the purpose is to show whether there's a close correspondence between percentage of POC in each district and the percentage of participants in the community engagement events who were also POC, a scatter plot may work better. This a quick makeover with a caption explaining how to read it:

Thursday, June 27, 2019

Visualization humor

Coincidences: After my post yesterday about a paper she co-wrote, Lace Padilla tweeted a cartoon she drew a while ago for her PhD dissertation:


I'm very interested in hurricane mapping—I've written about it for Eurostat in the context of visualizing uncertainty; there's almost an entire chapter about this topic in How Charts Lie; I'm part of a research group focusing on it; and if you've attended one of my recent public talks, you've heard me rant about it,—so I suggested a second panel for the cartoon, showcasing the infamous cone of uncertainty with dialogue that could go like this:

“Nice! The hurricane won't affect my town!” ”Oops, it still may threaten you.” ”Then why isn't my town inside the cone?”
. . .

Lace then drew an entire strip (below), adding: “This could be the start of "Everyone vs. Scientists" comic strip.” I disagree: it could be centered instead on visualizations, and the many ways the public misinterprets them. Maybe Lace herself or one of the many talented artists who design both pictorial infographics and data visualizations could create an XKCD-like cartoon series about graphics?

Wednesday, June 26, 2019

Our understanding of rainbow color schemes remains incomplete

Rainbow color schemes are the 3D pie charts of the sciences: they are are everywhere even if they are considered dubiously useful by visualization professionals. Last year Betsy Mason wrote a nice summary about the shortcomings of rainbow palettes for Scientific American magazine: rainbows don't just make you see categorical boundaries in otherwise continuous data, but also,
the relationship among the colors is not intuitive. “The problem with the rainbow is that you don’t perceptually see it as ordered,” says Colin Ware, a human perception and data visualization expert at the University of New Hampshire who was not involved in the study. “If you give people the colors red, blue, green and yellow, they will not know which order to put them in.” Another problem is the brain naturally interprets differences in brightness, or luminance, as representing depth, with the brightest colors at the peak.
This said, we may be just scratching the surface A recent paper by Sam Quinan, Lace PadillaSarah Creem-Regehr, and Miriah Meyer explores whether the illusory discretization that rainbow color schemes elicit is the same for all readers, for all data sets, and in all variations of these spectral palettes:

The answer is that, yes, rainbow color palettes are tricky, but (a) there are differences between them, (b) the nature of the data represented affects the way inexistent boundaries on continuous scales are perceived, (c) luminance doesn't seem to be the only factor that affects discretization; chroma and hue may also play a role.

The paper ends by wondering why, despite their repeatedly proven shortcomings, rainbow schemes are still so popular—convention, familiarity, and attractiveness, but also the fact that scientists find them useful sometimes,—and by suggesting paths for further research as, quoting the authors, “the visualization community’s current understanding of how rainbow color maps are perceived and used remains incomplete.”

Tuesday, June 25, 2019

Narrative patterns for data stories

Thanks to Maarten Lambrechts I've (re)discovered NAPA Cards, an initiative deriving from a workshop on data-driven storytelling in Dagstuhl, Germany. I wrote “re” because this project is discussed at length in Chapter 5 of the book Data-Driven Storytelling, which I read a while ago; I had forgotten about its website.

NAPA Cards is reminiscent of a classic article by Edward Segel and Jeff Heer, “Narrative Visualization: Telling Stories with Data”, and a follow-up comment by Philip Man, in the sense that it tries to come up with a taxonomy of data-driven stories. There's also this recent article by the data team at the Financial Times open-sourcing some story formats they commonly use.

The authors of the chapter in Data-Driven Storytelling focus instead on patterns, which they define like this:
A narrative pattern is a low-level narrative device that serves a specific intent. A pattern can be used individually or in combination with others to give form to a story.
Therefore, a single story can use multiple patterns based on “the data, the formal setting, or the particular audience and its assumed background knowledge.”

The choice of patterns depends on the intent of the designer: “Examples of narrative intents range from enlightening audiences, to evoking empathic response, to engaging them to take action, or to questioning their beliefs and behavior,” in addition to more specific ones such as “delivering convincing arguments backed with data, explaining a type or data, and sensitizing people to their existence and power, or simply educating an easily targeted online audience.”

I haven't been able to find the full chapter online, so I recommend that you get Data-Driven Storytelling, as the NAPA Cards website is much more useful when paired with it —and with the other chapters in the book, which cover topics such as ethics, interaction, and audience analysis.

Monday, June 24, 2019

An interview about How Charts Lie

SuperDataScience, a podcast hosted by Kirill Eremenko —who has done nearly 300 episodes already, some about data visualization,— has just published an interview we did a month ago. We talked about graphics in general, and about some examples from How Charts Lie.

The interview's page contains links to all resources mentioned, including book recommendations such as Leland Wilkinson's The Grammar of Graphics, Mike Monteiro's Ruined by Design, Meredith Broussard's Artificial Unintelligence, Cathy O'Neil's Weapons of Math Destruction, and Virginia Eubanks's Automating Inequality.

Friday, June 21, 2019

Sources, methodology, and limitations in visualization

Luís Melgar, a University of Miami news visualization graduate, is currently working at Guns & America (Twitter), an investigative reporting collaboration between ten newsrooms all over the U.S focusing on gun culture and its consequences. Luís, in collaboration with reporter Alana Wise, has just published a story that quantifies gun shots near Washington DC schools.

There are some nice graphs and maps in the piece —I copied a couple below,— but what I'd like to call your attention to is the section at the bottom, which links to sources, discusses the methodology they employed in a lot of detail, and discloses limitations.

If, as Nate Silver has suggested in articles and recent tweets, journalism is to become a bit more empirical, increasing transparency and accountability by adding methodology sections to stories, rather than relegating them to sidebars or external pages that few readers visit, is a good first step. I know, I know, other publications —FiveThirtyEight itself, ProPublica, and the like— are already doing it, but I wish it'll become standard practice in all news media.

Thursday, June 20, 2019

Visualization style guides

I'm a fan of visualization and infographics style guides. In the past decade I've put together a small collection that includes gems such as old ones from the Dallas Morning News and the Los Angeles Times, and more recent ones, such as the Urban Institute's. Urban's Jon Schwabish as a pretty comprehensive collection, as well.

The most recent addition to my collection —thanks to Rafa Höhr for the tip— is the graphics style guide from The London Datastore, the data portal from the city of London. Written by Mike Brondbjerg, the guide (PDF) provides sound advice on typography, color, and composition. As other documents of its kind, it's a good starting point if you're a beginner in data visualization and you're still unsure about style choices. Also, don't miss the Further Reading section at the bottom of the guide's website.

Friday, June 14, 2019

A Venn diagram matrix by FiveThirtyEight

Jan Willem Tulp has called the latest xenographic by FiveThirtyEight a “Venn diagram matrix.” I wish that'd stick as a name for this type of chart (read about the term “xenographic” here):

The chart displays total Twitter followers of Democratic candidates who have more than 500K followers (this is the bubble size,) and how little or how much the candidates' followings overlap . You can find it in a story by Gus Wezerek and Oliver Roeder. Here's what they did:
The people following candidates on Twitter are those who want to receive a steady stream of information about at least part of the 2020 campaign. Understanding how that tribe operates can tell us something about an influential slice of the electorate. So off our web-scraper went, dredging up every follower of the 20 Democratic presidential candidates who FiveThirtyEight considered “major” in early May, when we ran our script.1 The result was a data set with almost 20 million entries, which you can download on GitHub.
That's right; the data is available on GitHub, where the authors authors also wrote: “If you use this data and find anything interesting, please let us know. Send your projects to @guswez or @ollie.”

Thursday, June 13, 2019

Spatial thinking, abstract thinking, and visualization

There are many kinds of nonfiction books. Some feel like walking through a historical building following a fixed path with the help of a tour guide who points out whatever we should pay attention to. That type of book takes you from point A (not knowing) to point B (knowing more).

Other nonfiction books are playful and meandering. They seem not to be structured like linear narratives with a beginning and an end. They feel like if the aforementioned tour guide met you in the lobby of the building, opened all doors inside it, and told you: “I'll give you a brief introduction to understand where you are. After that, explore at will; you can find more information about the wondrous objects inside this building in labels next to them. Feel free to read some and ignore others.”

The goal of this type of book is not so much to teach; it's to inspire you and give you leads to new ideas, references, readings. If you remember the island of knowledge metaphor at the beginning of The Truthful Art, books of this kind may not expand your personal island, but they reveal promising directions to do so in the future. Maybe we should coin the term “hyperbook” to refer to them (after hyperlink).

Barbara Tversky's recent Mind in Motion: How Action Shapes Thought belongs to this second type. After a few chapters that sounded a bit too basic, I felt enthralled by the breadth of the book. I underlined many passages, took copious notes on the margins and on the first blank pages, and wrote down the titles of many papers and books mentioned. If you work in design, or if you're interested in the inner workings of the brain, Mind in Motion is for you —maybe not because of its content per se, but because of the many thoughts it'll rouse.

Here are some relevant passages (apologies in advance for not transcribing them, but I read print books.) The main point of Mind in Motion is that spatial thought is the foundation of abstract thought:

Much —if not all— thinking is action exerted over mental objects:

Because of how useful they are, spatial skills should be more broadly taught. This includes graphs, maps, diagrams, infographics, and the like, a point I've made in recent talks and that I suggest in How Charts Lie:

The book offers suggestions on how to get started with that educational program:

The first half of Mind in Motion ends with a reflection about the relationship between perception, imagination, and action. The more we perceive, the more we can imagine; the more we can imagine, the more we can do —and the more we do, the more we can perceive:

The second half of the book is even more relevant to visualization and explanation graphics designers. Here's a passage about the importance of context and purpose; Tversky doesn't mention the audience as part of that context, although I think it's implied that that's the case —how much or little you guess your audience knows about the topic of your graphic should influence the way you design it:

Some thoughts on the design of information graphics, ending with some elementary rules of thumb, which may sound familiar to some readers of this blog:

And comments about whether visualization conventions are always really just conventions, which is why it's important to think carefully before going against them:

Tversky also describes multiple experiments related to the effectiveness of graphics, explanation diagrams, maps, animations, and many other representations, and she extracts general design lessons from them. I'd get the book just because of these and the references related to them at the end. Enjoy.

Wednesday, June 12, 2019

The future of visualization lies beyond visualization

Jeff Heer has published the massive slide deck he prepared for a capstone talk at Eurovis, which you can watch here; there's also an related paper. Jeff has promised to write an article summarizing his main points. Here are a few personal takes, paraphrasing a bit:

(A) Visualization on its own isn't enough; it's always part of pipelines and processes. Therefore, it doesn't make sense to practice or study it in isolation. The future of visualization research and practice is in interdisciplinary synthesis, and “the practice of principled interdisciplinary thinking is our greatest asset”. Bravo:

(B) If we miss the focus on interdisciplinary collaborations visualization can go awry in different ways. The visualization process has many steps, and mistakes may appear in any of them. The challenge is that professionals from different backgrounds are capable of detecting problems in some of the steps below, but no one can detect problems in all steps if working alone. Quote: “We need analysis support tools & methodologies for end-to-end analysis, not siloed ‘statistics’ or ‘visualization’ tools”:

(C) Visualization has an accessibility problem: we need to explore sonification, physicalization, and other forms of encoding information. Being too late to them myself, they are areas I'm becoming increasingly interested in; remember TwoTone:

(D) Multimodality —using visuals, text, sound, touch, and so on, either simultaneously or supplementing each other— is a world to be explored: “given a formal visualization specification, how might we re-target a design to other modalities?”

(E) There are tons of known unknowns and unknowns unknowns in visualization. If you are a researcher, Jeff's slide deck is an inspiring and endless source of ideas.

(F) My favorite part: Jeff takes Richard Hamming's “the purpose of computing is insight, not numbers” and Ben Schneiderman's deservedly famous “the purpose of visualization is insight, not pictures” mantras and proposes a new one: “The ultimate subject of the visualization research community is people, not pictures.” I'd write “should be” instead of “is”, but I'd ask for an applause anyway: