Thursday, June 28, 2012

Graphics, uncertainty, and the flaw of averages

Many people I know who work in infographics and visualization have traditionally paid little attention to Howard Wainer. It's a shame, as the man is responsible for many fine and elegant books, such as Graphic Discovery and Visual Revelations, and for making William Playfair and Jacques Bertin available again for the lay reader. Quite relevant feats, I'd argue.

I bring Wainer to your attention because right after I finished sending The Functional Art to Pearson/Peachpit, I started looking for inspiration for its follow-up. Wainer is an author I go back every now and then for ideas. A copy of his Picturing the Uncertain World: How to Understand, Communicate, and Control Uncertainty through Graphical Display traveled with me and my family for vacation a few days ago. I am reading it for the second time (first time was when it was released) and enjoying it a lot.

What makes Picturing the Uncertain World so attractive is that it's made of short essays, most previously published in magazines like Chance and the American Statistician. As a consequence, the range of topics is really wide, although all chapters are related —as its subtitle suggests— to the visual display of information. The book has a bit of everything: rants against flawed displays, warnings against the dangers of relying on averages too much, discussions on little known maps by Charles Joseph Minard, and chilling pages about charts designed during the Holocaust. Wainer's style is delightful and witty. Here's a quote I underlined five minutes before sitting to write this post; it should be reminded to every hardcore fan of visual"simplification":

A key tool to understand uncertainty is graphic display (...). A display that shows the data in all their variability acts as a control, preventing us from drawing inferences based only on a single summary number (e.g., an average) that can only feebly characterize a complex situation. Using Sam Savage's delicious phrase, graphs can save us from being deceived by the "flaw of averages."

The funny thing is that I looked for Savage in Google and I found his book, The Flaw of Averages: Why We Underestimate Risk in the Face of Uncertainty. Its website has some informative (albeit a bit crude) animations. I love it when a great read leads you to discover other unknown gems.

Monday, June 25, 2012

Truth and beauty in infographics and visualizations

Moritz Stefaner. Photo by Decoded 2010 conference

I keep editing the Profiles section of The Functional Art and finding great insights by all the interviewees. Here you have a few fragments of my conversation with famous information designer Moritz Stefaner:

You define yourself as a “truth and beauty operator.” What does that mean?
In the past, I used to tell clients and family I worked as a “freelance information visualizer.” That has way too many syllables. Also, I was always forced to explain what I actually did, as people don’t really have much of a grasp on what it means. After I explained that I turn data into images and try to help people make sense of large data sets, I was often asked if I tended to prioritze truth or beauty, form or function: Do these compete in my projects? Do I have to trade off one against the other? For me, truth and beauty have always been two sides of the same coin. What I want to highlight in my title is that I don’t want to trade truth for beauty or beauty for truth: I work with both. These are the two things that I will evaluate design solutions against: Are we helping to find out the truth? Are we doing it in an elegant way?

So you cannot have truth without beauty, and vice versa.
I put truth and beauty at the same level. If you have only one without the other in a visualization project, you are not done yet.

Buckminster Fuller, the famous designer and systems theorist, said once that he didn’t think about beauty when he started a design, engineering, or architectural project. He was just concerned with its functions; he wanted to find the right way to devise the product. But then, in the end, if the solution he came up with was not beautiful, he knew something was wrong. For Buckminster Fuller, in some sense, beauty was an indicator of functionality and of truth.

For me, design is much more than mere decoration. Scientists and engineering focused people, the kind of professionals I usually work with, think of design as decorating a pre-existing structure. That is the wrong approach. Good design is tightly intertwined with the content it presents. It consists of thinking about what to show, what to leave out, what to highlight, how to structure information, what rhythm, visual flow, pace, you want your story to have, etc. That’s design. I would say that structure dictates pretty much what comes out visually.


There are designers, particularly traditional infographics artists, who contend that most data visualizations don’t really tell a story. They argue that contemporary visualizers basically throw readers into an ocean of data and they let them on their own to navigate without any guidance. Do you think this is true?
I think it’s true, and I think it’s positive and fantastic. The most fascinating thing about the rise of data visualization is exactly that anyone can explore all those large data sets without anyone telling us what the key insight is. We can look up our own hometown on a census map and look exactly what the data points are and not just the averages in the county or the state. In my work, I do not try to tell a story. I try to tell thousands of them. Obviously, you cannot present them all simultaneously or with the same priority, but still, they can all be there, hidden in the raw data, waiting to be discovered through the interface of your presentation. I think this is the key difference of the genre of data visualization compared to, let’s say, more traditional information graphics

Saturday, June 23, 2012

The difference between infographics artists and visual journalists

Why the job hunt is so hard, by The Washington Post. See it complete

While editing the interviews in the Profiles section of The Functional Art I came across several common themes. One of them was the role infographics desks should play in media organizations. In most newsrooms, designers are people who passively wait for reporters and editors to give them information to do infographics about. However, quite the opposite is true in the most successful departments. Here's a question I asked Steve Duenes, graphics director at The New York Times:

So you are not an “art” department. You don’t consider yourselves just “artists”.
Certainly, there is an “art” component to what we do, but we are not “graphic artists,” and we are not a service group. We want to eliminate the passivity that suggests we should style a dummy headline and wait for the real journalist to fill it with meaning. We want to report and present the content ourselves. Very often, we work in parallel with other news departments. We pursue stories in ways that are similar to the Metro desk or the National desk, tasking reporters and organizing ourselves to pursue information. This was the drill, for instance, right after the Virginia Tech shootings, in April 2007, and in response to many other breaking-news stories.

And here's Hannah Fairfield, about The Washington Post:

What changed in the newsroom when you were hired?
Years ago, the Post’s graphics department, like teams at most newspapers, was more service-oriented. Artists tended to wait for other desks to ask for graphics, charts, and maps, and bring information and possible ideas for layouts back to them. My vision was that the team needed more independence and autonomy in order to produce really creative, high-impact visual journalism. Based on my experience at The New York Times, I knew it could be done.

The graphics team was really enthusiastic about the change, but I had to win over some other parts of the newsroom. During the first few months, we spent a lot of time talking to editors and reporters from other departments, who brought very specific requests for the kind of infographics they wanted. “I really appreciate this idea you are bringing to us,” I said over and over. “But let’s talk about the story. We will read it, we will make a judgment, and I promise you we will provide the graphic that is right for it.”

I was lucky I had the full support from high-level editors, because this is a pretty profound culture shift. Before I was hired, I was transparent about the fact that I wanted to bring some sweeping changes to the desk: If they wanted service, instead of the kind of autonomy I wanted to implement, they could have hired someone else. 

How did you grow trust? The reason I am asking is that the reality in American newsrooms is not what designers in most other countries experience. In many Latin American and European newsrooms, for instance, infographics designers are second-class citizens, if you know what I mean.
You have to start with both sides. Inside the team, you can empower the infographics designers. Make very clear from the beginning that you expect them to be proactive, that the new standard includes a more engaged approach, and that you will fully support them.

And then, whenever an editor from a different desk comes to “order” a graphic, you say: “I will give this assignment to a person on the team that I completely trust. I know she will be able to deliver. You will not get what you ask for. You will get something that’s even better.”

So you have three people on that equation: the infographics artist, the traditional writer, editor or reporter, and a graphics leader serving as a sort of bridge—at least in a first phase, when everybody is learning what their new roles are. That is a model that could work in any newsroom.

Sunday, June 17, 2012

Infographics as moral acts

Edward Tufte is —among many other good things— the Oscar Wilde of information graphics and visualization: He tends to write in aphorisms and epigrams, so he is a very quotable essayist. Here's a paragraph of his that I hold dear:

"Making an evidence presentation is a moral act as well as an intellectual activity. To maintain standards of quality, relevance, and integrity for evidence, consumers of presentations should insist that presenters be held intellectually and ethically responsible for what they show and tell. Thus consuming a presentation is also an intellectual and moral activity." (Beautiful Evidence)

That paragraph speaks to the three persons that coexist in me: The journalist, the educator, and the information designer. Its main theme is central in The Functional Art: Correctly presenting data and phenomena in a graphic is not just a professional endeavor; it is also —above all— an ethical mandate. So is openly and candidly discuss mistakes, yours and others'. In infographics and visualization, the decisions we make on how to encode information, how to organize it, and how to present it, should be guided by a simple principle: Whatever improves citizens' interest in a relevant topic and their awareness and understanding of it is morally and ethically* good; whatever obscures the subject, trivializes it, or misleads audiences is bad.

There are cases where we can clearly identify conscious disregard for this principle, although they are rare: Fox News' many egregious (see pie chart below) and epically flawed charts should be a tempting research topic for PhD students.

All media organizations commit blurrier sins every now and then, though. The best out there sometimes publish graphics that I'd argue are ethically questionable even if designers made them with noble goals in mind. Let's mention The Economist's occasional love for truncated axes in bar charts (needless to say, a bar chart baseline should always be 0), or this op-ed by Charles M. Blow, addressing inequality and unemployment in the U.S. As many of Blow's articles, this one must be praised for its reliance on empirical evidence, something that is not common among opinion writers anywhere. Unfortunately, when it comes to charting, we find this (see it complete here):

I'd argue against this graphic by evoking an old adage: If precise comparisons are crucial, circles are not a good choice. Can you really see that a person in the top income quintile would save ten times more than an individual in the middle quintile if income taxes were reduced by 20%? If you are absent minded, as I am, you'd walk away from this chart thinking that differences in savings are much less worrying than they really are.

You may contend: Well, you can still read the figures and visualize the real differences. And my answer would be: True but, then, if you want me —your reader— to mentally compare, what do you want the circles for? Just design a table.

A very recent case of misdeed is this unfortunate series of maps about the European economic crisis. They are among the best examples I've seen of how not to design cartograms. Let's take just a small portion of the graphic (above): Without reading the numbers, quickly tell me if the unemployment rate is higher in Spain or in Portugal. There you go: Portugal looks like the worst offender in Europe, when it is doing better than Spain and Greece...

(The ethics of infographics has been a concern of mine for many years. Read this manifesto at the Harvard Nieman Watchdog site to learn more.)

I am using the terms morals and ethics in the sense explained here.

Thursday, June 14, 2012

Embellishments (sometimes) have a place in information graphics and visualization

Maybe because of the title of my book and some posts and articles, like the previous one, some people assume that embellishments should be systematically rejected in information graphics. They think —I know it because I've been told so more than once— that being and advocate for graphics that inform and respect the reader's intelligence and what we know about visual perception makes you a promoter of ugliness, dryness, and boredom.

This is far from truth. I do love beautiful infographics and I have the feeling that a small part of what is usually considered chartjunk (a term coined by Edward Tufte) is not really junk. Smart pictograms, unobtrusive icons, elegant and subtle visual effects (see below), etc., may sometimes have their place in charts, maps, and diagrams, depending on the publication and its audience (let's bring up Otto Neurath and Gerd Arntz).

The reason a few ornaments can be appropriate is that, when correctly used, they may enhance memorability: They might help you remember the graphic and its content in the future. Besides, as Donald A. Norman said in one of his books, a small amount of useless beauty can even make an object more usable. (I want to be extremely careful with my wording here: I am not making the case for things like this or this.) Here's the catch: decoration, if added at all, must not detract from the information, must not hurt readability, must not become the center of attention, and must not take away space that should be used to better explain your story. You should not give up depth because of aesthetics. Ornaments should be like headlines in news stories: they grab your attention and predispose you to explore relevant information. They put you in a good mood.

The three paragraphs above are just an excuse to link to a visualization that has caught my eye this week, the latest wonder by The New York Times graphics desk: "Where the Heat and the Thunder Hit Their Shots". Are the animated transitions between tiled photographs and heat maps useful at all? Maybe not —they don't add anything that a small mug shot on the corner of each map would not achieve,— but they certainly make the graphic memorable.

Tuesday, June 12, 2012

Random thoughts on infographics, simplification, and the revolting 10 seconds rule

In what sense is this infographic "easy" to read, exactly?

A few random thoughts: When surfing the Web in search for good and bad information graphics and visualization examples, it is not hard to feel sympathy for Gizmodo's Jesus Diaz's rant, published a while ago. Just try to understand the pie chart above and evoke with me:

"The fact is that these monstrosities are not infographics. These atrocities are crimes against good taste and everything that infographics really should be. They're just a bunch of statistics jammed together on horrible vertical pages, bloated with bad drawings and clipart created by primitive monkeys using CorelDRAW! 1.0—graphical disgraces that most often disguise spam, commissioned by iniquitous companies looking to increase traffic to their sites."

It's been a while since marketing and PR agencies started taking over the very word "infographic", which twenty years ago was mostly used to refer to charts, maps, and diagrams in newspapers and magazines. Nowadays, if you go to Google and search for "best infographics", you will find endless collections of the horrors Diaz mentioned. Many of them are bad not because they look terrible (some of them do, of course, others don't) but because they don't live up to their own hype. According to this article, for instance, "more than any other method, infographics convey data quickly and attractively." That's certainly not the case of the examples there, which are muddled, busy, and difficult to read. 
Besides, that same piece reminded me of some managing editors I've had to deal with in the past, as it echoes a sentiment I've come to profoundly despise:

For a good visual to work, it needs to get its message across in ten seconds or less."

Well, it does, but only if your infographic is going to be used as a promotional tool 
—an ad— and not as a visual display of evidence. The problem with this kind of articles is that too many managers in news media (who should know better, but don't) tend to equate marketing infographics with traditional infographics: they go to, see some colorful and fun compositions, and ask their designers to mimic their style: cute, ininteligible, and shallow stuff that can be "easily digested", gorged as fast-food, and not savored and enjoyed thoroughly and slowly, as good meals should be. To fight this way of thinking, let's remember what William Cleveland said in one of his classics:

"The important criterion for a graph is not simply how fast we can see a result; rather it is whether through the use of the graph we can see something that would have been harder to see otherwise or that could not have been seen at all."

Cleveland was writing about scientific visualizations and graphs, but his words can be applied to any area: while it is true that a good graphic should attract the audience's attention, the same way a good headline does, its superficial beauty should be just an entry point to a deep discussion of the topic at hand. Most of the infographics we see online today are the equivalent of writing a news story that offers just a headline and a short lead, and nothing else after that: no context, no background, no opportunities to delve into the information.

Some of the agencies that advertise their
infographics services are hilarious pools of self-contradiction. There are usually great mismatches between what they say they do and what they actually do. Take this one (PDF):

Here's how the company describes its goals:

"Meidata provides senior managers with accurate business information to support decision making process. 
One of Meidata's advantages is the advanced visual presentations of analysis findings and data. The unique infographics of Meidata assists senior managers by enabling clear presentation of large amounts of information. Our infographics save time for senior managers and assists in conducting clear messages through the organization."

It may well be that this poster looks nice on a wall but, analysis? What kind of insight does that display provide when some of the forms chosen to encode the data go against efficiency standards? 
(Namely, if you want to compare with precision, don't use circles.) Would you want to work for a company whose senior management uses such a thing to analyze its data? I wouldn't.

Regarding the "simplification" information graphics apparently
facilitate, and that I discuss (critically) in The Functional Art, here's a quote from an amazing essay by Reif Larsen:

"A large part of the infographic’s intrinsic appeal seems to lie in its visual reductionism of complex information. Reductionism itself is not inherently bad—in fact, it’s an essential part of any kind of synthesis, be it mapmaking, journalism, particle physics, or statistical analysis. The problem arises when the act of reduction—in this case rendering data into an aesthetically elegant graphic—actually begins to unintentionally oversimplify, obscure, or warp the author’s intended narrative, instead of bringing it into focus."

As Nigel Holmes said once: information graphics should not simplify; they should clarify. There's a huge difference between those two words. Infographics 
don't "illustrate information that would be unwieldy or extremely boring in text form" (reference). If you think like that, you should choose a different career. What infographics must do is to allow you to see beyond what bare eyes can see, reveal trends and patterns, let you gaze the invisible, in general. That's what they are for.

Sunday, June 10, 2012

An information graphics and visualization reading list

If you are going to collect books, you better hire reliable bookkeepers

Every now and then I find a message in my Inbox asking for an information graphics and visualization reading list. As my students know, I love to recommend wonderful books. So here we go.
(Notes: (A) If I were you, I'd take a look at these books more or less in the order they are presented, as the list starts from the basics and goes to the more advanced stuff; (B) I don't suggest any book about software; (C) I don't mention all books about graphics that I think are relevant see the photograph above; there are many more,— but just a small and maybe a bit idiosyncratic selection that can help you get started; (D) I have not included The Functional Art for obvious reasons!)
1. The Back of the Napkin, by Dan Roam (Amazon)
Roam shows how to visually organize and display information using just pen and paper. This book is fun and includes tons of hand-drawn examples.

2. Nigel Holmes on Information Design, by Steven Heller (Amazon)
The art director of The New York Times interviews one of the best infographics designers ever. 140 pages of insight.

3. Visual Language for Designers, by Connie Malamed (Amazon)
Malamed discusses the psychology of visual perception and cognition and offers good examples of information graphics.

4. Show me the Numbers, by Stephen Few (Amazon)
One of the best books about data and charts I know. See the post linked above to learn more.

5. How to Lie With Maps, by Mark Monmonier (Amazon)
Cartographic concepts may be hard to grasp if you try to read an advanced textbook before going through the basics, so first read this shork book. When you finish, projection, scale, and symbolization will not be mysteries anymore.
6. Malofiej book series (Amazon)
These are the books to get if you need some inspiration. They collect the annual Malofiej Awards, organized since 1992 by the Spanish Chapter of the Society for News Design. The text is bilingual (Spanish and English) and, besides tons of great graphics, these volumes include articles by leading professionals.

7. Creating More Effective Graphs, by Naomi B. Robbins (
A catalogue of good and bad practices in statistical graphics design. Very informative and well organized.

8. The Wall Street Journal Guide to Information Graphics, by Dona M. Wong (
An elegant and beautifully illustrated short guide to charts, written with the news designer in mind.

9. Visualize This!, by Nathan Yau (Amazon)
If you've heard about the FlowingData blog before, I don't need to introduce you to Nathan Yau. His book includes a short intro to graphics and then plenty of useful details about programming languages.

10. Thematic Cartography and Geographic Visualization, by Terry A. Slocum (
This is the book I like the most in my small cartography collection, and it is probably the one from which I have learned the most about mapping. If you still don't know what a "choropleth" map is, then run to the bookstore.

11. Information Visualization, by Colin Ware (Amazon)
This impressive book was life-changing for me. Thanks to it, many things I did everyday at the newsroom suddenly made a lot of sense, while others stopped making sense at all.

12. How Maps Work, by Alan M. MacEachren (
If you are following the reading path I suggest, before you get to this massive volume you may have gone through Ware's, so you will be ready for the oceans of fascinating information in MacEachren's pages. This is the most comprehensive explanation of why our brain is wired to understand visual spatial representations.

13. The Elements of Graphing Data, by William Cleveland (
A spartan-looking classic that may be hard to digest if you don't read Tufte, Few, and Robbins first. It is a must-read if you are serious about your career, though.

See the column on the right for more readings suggestions.

Thursday, June 7, 2012

A promo video for "The Functional Art"

This video will show up soon in the The Functional Art pre-order page in
I hope you like it.

A new edition of Stephen Few's "Show Me the Numbers"

The front cover for the 2nd Ed. of Show Me the Numbers

Like many others in this business, I ended up designing information graphics and visualizations for a living with no prior formal training. As I explain in The Functional Art, I attended Journalism school in Spain and, due to a chain of coincidences, I got an internship in a very good infographics department, at a newspaper called La Voz de Galicia, where I learned by doing and by watching.

When I moved to UNC-Chapel Hill in 2005 to teach information graphics to journalism and design students, I was in the awkward position of not being able to properly explain why things were done the way they were done: Why would you choose a bar chart instead of a time-series plot in this particular graphic? Why are pie charts not the best solution in most cases? What would you prefer a subdued color palette instead of one based on pure reds, blues, greens? I did have intuitive answers to those questions, but I needed to somehow find scientific and logical foundations for them.

From that point on, I forced myself to follow a strict reading regime: I basically put my hands (and eyes) on every single book and paper I found on statistical representation, cartography, information design, and cognitive science. The Functional Art is, actually, a synthesis of everything I've read, combined with what I've learned during my career as a designer and visual journalist. Think of Tufte, Kosslyn, HolmesWare, Cleveland, Robbins, MacEachren, Monmonier, Robinson, Spence, WainerShneidermanHarris, Koch...

I also read Stephen Few's three books: Show Me the Numbers (first edition), Now You See It, and Information Dashboard Design. Together, they are the best hands-on introduction to the use of graphs for data analysis and presentation I know. But it was the first one, Show Me the Numbers, that influenced my own approach to graphics the most. Even if Few specializes in Business Intelligence, not in design or journalism, and he focuses just on graphs, not on information graphics and visualization in general, his ideas on clarity, concision, honesty, and respect for the data and facts are applicable to any area related to the visual display of information. If you've ever thought that Tufte's four are too abstract and difficult to translate to the real world, try Few's.

Considering how highly I regard Show Me the Numbers, knowing that a second edition has just been released was great news. I had the opportunity to read it in advance and I can say this is not just a minor refurbishing of old material, but a completely new book: It includes nearly 100 extra pages, new chapters on bad graphs and data narratives (telling stories with numbers), and appendices about Excel and color palettes. The first edition was a must-read. This new one goes beyond that.

Sunday, June 3, 2012

A fun pie chart about pie charts

One of the recurring jokes among my students at the University of Miami and at UNC-Chapel Hill is that I hate pie charts. I don't. Actually, if in a project you prove to me that a pie chart is more efficient than any other kind of graphic at communicating your story, it's likely you will get an A.

Circular graphics —pie charts, comparisons based on bubbles, spider charts, and the sort— have their space in information graphics, but they are so often misused that it has become a sport for experts to criticize the most egregious examples of wrongdoing (see a summary here) with wit and humor. I sympathize with the view advanced in those articles, in general, with just a few exceptions: in the The Functional Art DVD (you'll get it with the book), I praise this circular plot by The Guardian for very specific reasons. Stephen Few would not approve; see his classic "Our Irresistible Fascination with All Things Circular."

Anyway, I am getting to the point. See how people sometimes make fun of my distaste for pie charts: In September 2011, my friend Matt Perry, head of infographics at the San Diego Union-Tribune, published this picture on Twitter; its caption was "Donut social time in the graphics department."

It was a slow Friday for me, so I replied, tongue-in-cheek:

Here's Matt again:

And here's me, trying to sound smart:

Only to be effortlessly outwitted by Matt:

Who attached this graphic to his tweet. He produced it in five or ten minutes. I am using it in The Functional Art DVD, by the way. How could I leave it aside? It's so funny:

Friday, June 1, 2012

Stefanie Posavec: Infographics and Literature

I wrote about Stefanie Posavec right after I talked with her on Skype a few months ago. I decided to include her as the last interviewee in The Functional Art because —even if she is not a journalist or an information designer (she calls herself a "data illustrator")— I am a devoted fan of her elegant and intricate literary visualizations, particularly of her Literary Organism, which is one of the graphics that we discussed during our conversation. Click to enlarge:

The Literary Organism, which is part of a series of visualizations that I showcase in The Functional Art, displays the structure of Part One of On the Road, by Jack Kerouac. It visualizes it using a tree structure. Part One divides into chapters, chapters divide into paragraphs, paragraphs divide into sentences, and sentences divide into words. Everything is color-coded according to key themes in the book. Here you have some excerpts from the interview, where Stefanie explains the project. What is more impressive to me is that she did all this manually, using just color markers and Illustrator:

How did you design the Literary Organism? Did you use any script to count the words, organize them, sort them according to themes, etc.?
Believe it or not, I didn’t. I did it all by hand. I could not get an electronic version of On the Road and I couldn’t figure out how to digitize the copy I had, either. So I ended up counting all the words one by one, and sorting them by key themes. I spent a lot of time going through the book and highlighting different sections (see images below).

I made the graphics in Illustrator. They were not generated with code. I know it’s possible to do it using scripting, and there are wonderful tooks for that, like Processing, but I don’t know how to use them.  I am aware that I need to automate, but sometimes I feel that it’s important to spend that kind of time gathering your information by hand. It feels a little more natural. Also, it creates bonds with what you are working on: I had to read the On the Road over and over again, so the outcome was as much a representation of the text as it is a representation of the novel in my head, of my experience of exploring it.

Did you want to make a particular point with your graphics or did you just wish to create an art project?
To tell you the truth, my only goal was to be able to put the entirety of the book on the wall, maybe just with the intention of inspiring awe and wonder: after seeing the graphic, I wanted people to see On the Road in a different way.

As its name says, I consciously made the Literary Organism feel organic: I am intrigued by the parallel between how books are celular, and plants and animals are celular as well. I wanted to make that connection. With the circular arrangement of lines and colors I also tried to convey the rythm of the book. On the Road reads like a poem. So the graphic is not so much about insight, but about making you think differently about the book.

Here you have a photograph of Stefanie's copy of On the Road, and a selection of its pages, covered with highlights.