The expertise, not the pedagogy

Today I attended Edward Tufte’s one day course on presenting data and information.  It involved sitting in a double wide airport hotel ballroom from 9am to 4pm; the first hour consisted of a “study hall” in which I was given four (beautiful) texts and a list of assigned reading.  Lecture started at 10a and continued until 1:15p. There is one 10 minute break at 11:30a.  After an hour lunch break, lecture resumed from 2:15p until 4p.


In terms of offering foundations and inspiration around beautiful data design, Tufte is (obviously) a great subject matter expert. He’s not a bad storyteller, either. Here are some takeaways from the five hours of lectures.

Open mind, not empty head. Analytical thinking is critical to consumption and creation of analytical design. Understand the assumptions and motivations of the data source. Consider the likelihood of accuracy (e.g. are they trying to measure something really difficult to measure?). Consider the potential relationships between information and conclusions. Consider the endpoints: if you can’t model an endpoint from the graph as provided, then something is most definitely wrong.

Content and credentials. A visual is only as good as the content presented and your impact is only as good as your documentation.  Build your reputation for consistent quality.  Include not only the name of your organization but the individuals who created the graphic.  Include links to your raw data.  Make sure the data is interpretable by others.

Always present a flat surface.  Avoid the slide deck.  Avoid the slow reveal.  Don’t try to control the order in which people read things because they are going to prioritize differently or find different things interesting. Humans understand more, see more, create more when they have access to spatial adjacency (My advisor used to call this “screen real estate”).  They also love to scroll.

Get out of your own voice.  Quote the experts who can say things in colorful ways that you cannot.

Data paragraph as your unit of analysis. The most effective combination of visuals, words, and numbers is what you want to create.  Look at ESPN.  Look at the New York Times. National Weather Service. These sites do it well.

Limit the non-data ink. Grid marks?  Only if it helps the consumer and then make them as simple and light as possible. Hyphens, colons, and other repetitive administrative considerations should be deleted. Duplicated data (in most cases) should be erased.

Direct labeling. Good design increases content reaction time and format decryption time. Get rid of “one-off” coding – that is, codes that consumers have to comprehend for just one graphic. Get rid of the legends and label the map.  If you have data that must be labeled in very specific ways, write it in a sentence.

Present. In sentences. I will say it again (Edwin Tufte certainly did): Sentences are one of the best forms of direct labeling.  No bullet points.  No sentence fragments. Present data in sentences, even if they aren’t presented in paragraphs.

Annotate (with sentences). All lines should be annotated by actions, causes or events…preferably in sentence form. While traditional organizational charts and hierarchies tend to focus on nouns, the real world is networked, a complex combination of interactions and feedback. Everything in this world is about interaction. There were direct shout-outs to Tim Berners-Lee, but (surprisingly) nothing related directly to Bush, Englebart, Xanadu, or annotation systems.

Multivariate. Multimodal. Multilayered. Use the design to highlight contrasts, comparisons, causality, correlation: an analytical design is always the result of sound analysis of content.  Find ways to reveal different aspects of the same topic – incorporate information about sight, sound, behavior, context.  Finally, layer data in ways that can peacefully co-exist.  Google Maps layers light color in ways that does not obscure: interstates, streets, places of interest, natural landmarks…all on top of an aerial map.

Openness. Openness was an underlying value. Open access, source, data were frequently present, as were references to the open web and the importance of translational research. Tufte deals in an impressive amount of self-promoting retail in the lobby, but he’s not all marketing and hype. I give him credit for telling us how to get around paywalls for research articles. Also, I saw people taking pictures of his slides during the presentation – none of the earpieced men stationed around the ballroom tried to stop them.


I have difficulty categorizing this experience in terms of my pedagogical expertise. It wasn’t a workshop because there wasn’t a shred of experiential learning included. No moments of reflection.  No consideration of knowledge transfer or application. It wasn’t a seminar because there were no opportunities for interaction, no social learning.  No small groups.  No co-constructed knowledge.  Not even a single question and answer period. Finally, it’s not really a course because there was no assessment, no learning projects, no evaluation, no credentials, and it only lasted for five hours.

Can a five-hour lecture which has no learning assessments or learning products really be called a course?

And I say firmly, no.

Additionally, Tufte had a really interesting approach to people.  Here it is, as best as I can convey it, paraphased and synthesized over the lectures:

Always treat your audience with the utmost respect, because they are successful in reading high level data every day, as evidenced by the success of and  However, they will not always be right, nor will they always do the right thing.  Sometimes you must force your audience to sit with the big picture for a while.  You must force them to be patient. It will be better for their overall understanding than allowing them to drill down, zoom in, click the link, get to what they want.  Finally, understanding their reaction to your work isn’t necessary for excellence; indeed, feedback is great for finding problems or gaps, but excellence is an entirely different mindset.

I’m not 100% opposed to this argument.  Respect is good. People aren’t necessarily right. Requiring time with the big picture isn’t a bad strategy.  Understanding the difference between “acceptable to your audience” and “innovative beyond your audience’s wildest dreams” is important.

However, Tufte’s approach to “respect” is one-size-fits-all, and it is his lens of how respect might appear. I heard him mention color vision deficiencies only once.  There was no mention of ADA guidelines.  No discussion of screen readers, alt text.  No mention of those with dyslexia or who may not read well or may not read English.  There were only two mentions of the fact that his left-to-right designs were built for the Western world. Sometimes your audience is not reading or  Sometimes your audience doesn’t even have access to the Internet.  Sometimes they are just trying to understand a cancer diagnosis or their bills. I wonder how Tufte feels about these individuals as audience?

I also take issue to Tufte’s pedagogical orientation of “audience.” This was a sage on a stage situation and I don’t think it’s that way just because of the setting (traditional lecture room set up in a double wide ballroom). This is a man who reveres his own authority.

Did I learn something today? Obviously the answer is yes (see above) However, I did so because I made my own active learning experience.  I made my own backchannel with a friend – we were emailing about what I was learning throughout the day.   I made the decision to report out via this blog post.  I made myself come up with applications for my job and my personal life.  If I had not done these things – if I had not put the trouble into writing this up…how much of it would I remember?  Not much, I don’t think.

Edward Tufte lives and dies on his main argument: the data visualization is only as good as the content.  His argument is incomplete.  Data visualization and this course is only as good as people’s understanding, application, and remembering of the content. In education, you need more than paper-based, slide-based, or even multimodal content. People are content (thank you, Dave Cormier). Process is content.  Edward Tufte needs to wrap his head around a different definition of content.

And why do people just sit there and take it? Because they don’t know better.

So.  Go to the ‘course’, but do it for the books, not the pedagogy.

Featured Image: Brandi Redd

Edit (6/6/17):  Thank you to my dear friend Ken Bauer, who pointed out that Edward Tufte is Edward and not Edwin – I corrected throughout.  In my defense, Dr. Tufte never once introduced himself throughout the entire 5 hours of lecturing, so I’m not 100% sure it was him at all :).

One Comment Add yours

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s