Hacker Newsnew | comments | show | ask | jobs | submit login
The Financial Times on Edward Tufte (ft.com)
113 points by DanielRibeiro 805 days ago | 39 comments

In the five years that I've practiced data visualization and analysis, the day-long Tufte seminar that my boss sent me to was likely the most important single day of my development career. Tufte has virtually nothing to say about how visualization should be done off paper, but his general principles are invaluable. I was still making annoying Flash charts at the time, shortly after attending Tufte's workshop, I realized how much of an anti-pattern it is to rely heavily on button-pushing and on-hovers.

It's true that Tufte's pioneering work was done in the age of paper, but that simplicity and clarity of purpose is even in greater need today.

(And I'll also say, Jesus Christ this submitted headline made me fear that he was dead)

Edit: I can't avoid saying that some of the more contemporary ideas he's really into are sometimes a little boring compared to his data visualization ideas...if you live in New York, Tufte has a gallery of art that he's done and space devoted to others' ideas...I thought the documentary on the designer mentioned in the OP was a little dry. Also, at his gallery, Tufte has a short video on loop about how he'd improve the Apple iPhone weather app...I forgot the details but I remember thinking that I disagreed with a lot of what he proposed.

(On certain days of the week, Tufte is actually in the gallery...he was very nice to talk to)

His gallery: http://etmodern.com/ETmodern/ET_Modern.html


The world would be a better place if we all embraced Tufte's driving principle of respecting people's intelligence, rather than assuming stupidity.

Going forward, Bret Victor's work does an utterly fantastic job of blending Tufte's principles with 21st century interactive technology.


Tufte agrees. https://twitter.com/EdwardTufte/status/336951232197451777


> It's true that Tufte's pioneering work was done in the age of paper, but that simplicity and clarity of purpose is even in greater need today.

His principles aren't infallible. This 2007 study found that a group of students preferred a graph with "chart junk" to Tufte's minimalist style, although the study does note that this may have been caused by the students' familiarity with "chart junk" graphs.



Eh. I haven't read the study and it's behind a pay wall, but from your description and the abstract this sounds like it's measuring subject preference.

Whether subjects prefer "chart junk" is orthogonal to the question of information comprehension and retention, though.


The paper you link to is pay wall protected so I'm loathe to dismiss it without actually reading it...but the fact that it's been cited once and has this in its abstract:

> Motivation -- To evaluate people's acceptance of the minimalist approach to information visualization.

Um, OK...there are so many things wrong with that as a scientific hypothesis that I don't even know where to start. Perhaps the fact that this hypothesis was tested against a single example is a sufficient red flag? Doesn't it feel a bit of a stretch to refer to this study's findings so matter of factly?

In any case, though, I agree with this study. And so maybe would Tufte...it's not about what people prefer, it's about what people actually learn and understand from the graphics. Is your purpose to clarify and educate? Or to please the eye? I don't think Tufte would ever argue that he would win in a contest of the latter (it is true, though, that elegance often correlates with simplicity)...so what this "study" is trying to prove entirely misses the point.

Do a study of whether an average group of people are more drawn to the works of Ernest Hemingway versus Dan Brown...I don't think you'd find the results of that study being a conclusive argument that Dan Brown is among our greatest writers.


The study was scientific in measuring subject preference. That doesn't say much obviously. Many people prefer hamburgers to salads, and eye candy is often preferred even when it reduces efficiency.


There's a semi-recent study showing that people's recollection of the material two weeks later is higher when they see a chart junk version vs. the clean chart. That's the real game changer imho.


I suspect that this is a far more meaningful paper to link to: http://dl.acm.org/citation.cfm?id=1753716

I declined to go through the paywall, though, so I have not read either paper.


Well that nearly gave me a heart attack. I thought for a moment that this was an obituary. Hope he sticks with us for a long time.


My reaction too. That's a rather odd way of presenting an article.


Yes, a very poorly chosen title.


Edward Tufte says Bret Victor [1] is someone to watch out for. I am curious: can someone who is into usability look at the site and comment on whether this is considered usable? I mean his site is incredibly innovative and beautiful. However, I find it kind of hard to navigate?

[1] http://worrydream.com/


It's not about the website, it's about the content. If you read the articles, I think you'll be suitably blown away. Example: http://worrydream.com/#!2/LadderOfAbstraction


Oh no, I am impressed. The guy is as close as one can come to be a modern renaissance man. The website is also a statement of art and I was curious how it came across from a usability perspective.


Watch his presentations, they are incredibly inspiring and impressive.

https://vimeo.com/36579366 https://vimeo.com/64895205 https://vimeo.com/66085662

All are required viewing for the HN crowd.


Bret Victor, who has been inspired[1] by Doug Engelbart, has showcased plenty of innovative ways on design/user interaction on his presentations/writtings:

* Magic Ink[2]

* Stop Drawing Dead Fish[3]

* Drawing Dynamic Visualizations[4]

* Media for Thinking the Unthinkable[5]

* LearnableProgramming[6]

* Inventing on Principle[7]

[1] http://worrydream.com/Engelbart/

[2] http://worrydream.com/#!/MagicInk

[3] http://vimeo.com/64895205

[4] http://vimeo.com/66085662

[5] http://vimeo.com/67076984

[6] http://worrydream.com/LearnableProgramming/

[7] http://vimeo.com/36579366


I'm sad that Apple took a couple of years of Bret's innovation and locked it away. Thanks, Apple.

I couldn't agree more with ET's two choices. Mike Bostock is a machine. He's produced a lifetime's worth of awesome work, and he's just getting started.


I've put in a request to my boss to attend his one day seminar last year. However, recent reviews of attendees say his views haven't caught up with modern times and he just sticks to visualization techniques from years ago.


I would agree with that characterization. He didn't present principles of design, or anything you could walk away with and apply. He showed various examples of design, said what he liked about them, intermingling snarky, off-the-cuff remarks. (He hates Microsoft, as he pointed out multiple times.) In addition, there was a lot of personal anecdotes that felt like self-agrandizement.

What was good about it was that an occasional comment was thought-provoking. And the price of the seminar was not a whole lot more than the retail price of his books, which were given to all attendees. So it felt like a good return on the investment. But in terms of pure instruction, it was indeed disappointing.


I recommend getting his books, or just The Visual Display of Information (http://www.edwardtufte.com/tufte/books_vdqi) if you don't want to go a seminar.


Just to put in a contrary opinion, I was really disappointed in this book and I don't understand the praise it gets. Here's my extensive review on Amazon: http://www.amazon.com/review/R11NYC3OE3LBE


Personally, I came to very different conclusions. I found the book thought-provoking, and while your criticisms seem to be factually accurate, so was the original material you were criticising in each case I checked.

I am upvoting you anyway, because we should never take things for granted just because someone famous said them, and because I appreciate that you took the time to make a careful and reasonable argument even if I personally disagree with it.


I am really curious to know what you took from this book. I am really confused about why people like it so much.

For example, do you like Tufte's box plot? Do you like his revised bar chart? Do you strive, in your own graphs, to "erase redundant data-ink" in the way that Tufte advocates?


I read the book not long after it was first published and relatively early in my career.

What I found useful was not the spartan presentation. I do a lot of custom UI and data visualisation work these days, but even my plainest designs don't tend to be as extreme as many of Tufte's examples.

However, numerous projects I've worked on have been influenced by his underlying themes, like avoiding misleading presentation and unnecessary clutter, using graphics to convey patterns and relationships that wouldn't stand out from the raw data, and most of all how effective a customised or completely original presentation style can be depending on the nature of the data and what you're trying to illustrate.

Maybe if I read the book for the first time today I'd regard a lot of the ideas as obvious or trivial, just as I don't need to read another article on different types of colour scheme or how to choose well-matched fonts any more. For on-line use, where we have tools like colour and animation and interaction readily available, I certainly do things very differently. However, a lot of the way I look at data visualisation and graphical illustration today was no doubt shaped by reading Tufte early on, and so I give him the credit I feel is due.


I attended a seminar a few years back. The books are great, but I actually really enjoyed his commentary on the designs. He really explains well why certain designs work well for a given purpose and fail miserably at other purposes.

For example, the Minard chart[1] works exceptionally well to convey the loss of life in Napoleon's campaign in Russia. However, if you may not be able to apply this type of chart to a typical dashboard as easily. Tufte does a great job explaining exactly what makes Minard's visualization work so well.



Many things he shows are 'clever' and have a high density of meaning, but it doesn't mean it's intuitive.

Here's an admittedly elitist view: If _We_, HN readers, think its clever or neat, it's probably over the heads of the average audience.


I went to one of his seminars last year and I really enjoyed it. Yes, some of the material is old but he did a section on how you should get rid of powerpoint and start using slides that look like webpages.

People understand how webpages work and can quickly grab the information they want and it doesn't lull them to sleep like a powerpoint. He also said that baseball box scores on ESPN might be one of the best data representations. After you've seen one box score, you automatically get it from then on out.


i did the day long seminar a few years ago. it's interesting and engaging to watch, but it's mostly alot of hot air.


I attended one of his day long seminars recently, and I was displeased with the lack of breadth and depth. Sure, he rails against PowerPoints like no other, we get that. My biggest issue is that everything he presented was from someone else-- he's literally just aggregating other people's work and postulating on them with fuzzy emotional thoughts-- and barely anything applies to dynamic, user driven displays such as you would find on a web dashboard.

How to save yourself $300 hundred dollars AKA My half page of notes from the seminar:

0) Get the books on eBay or Amazon if you _really_ want them.

1) Don't use PowerPoint. Try to convince your viewers to browse a web page as the presentation, after you've convinced your IT staff to set up a web server for your presentation.

2) Look at ESPN.com, Weather.gov and Google News and Google Maps if you want to see what he thinks are the pinnacle of data presentations on the web. This will save you a few hours.

3) "Use verbs not Nouns" Show comparisons and mechanisms.

4) "Do whatever it takes to express the information"

5) "Show interaction between objects"

6) "Don't use boxes or drop shadows". Avoid "Chart Junk"

7) "Use small multiples" (he peddled his spark lines and ESPN's groups of stats)

8) Make paper based, text presentations and supply that to people as the arrive at the meeting. This way they can read that before jumping in with questions. The paper allows them to skip ahead and read what they want to know about so they don't interrupt you.

9) Calm the f2ck down when you go to meetings. He spent about an hour giving a lame rendition of How to Win Friends and Influence People. In one sentence: Be nice, be patient, be persuasive.

10) Ask yourself:

==1) "What's the problem?"

==2) "Why does it matter?"

==3) "What are you going to do about it?"

Other issues I had with his seminar:

1) I disliked the way he was surrounded by his thug ushers, who seemed to comb the aisles looking for someone or some thing. It made me feel like I was at a cult meeting or that they were worried about someone recording the seminar.

2) I disliked his peddling of his mother's book and his sculptures.

3) His dour emotional expressions about not wanting to teach government how to visualize things (This meeting was in Arlington VA, so I'm sure there were lots of military folks/contractors) in wake of the leaks. That's fine but now you look like a sellout, 'teaching' them anyway.

4) The multiple videos presentations about visualizing musical data. That's cute, but the guy is off his rocker if he thinks people are going to use that in place of a traditional presentation.

5) The repeated videos. I think we watched the same video three times of a Swiss topological map with an ethnic female voice narrating. By the second time I was already thinking to myself 'I wish he would stop wasting my time with that'.

6) I understand there is a cult of personality around him, and I generally thought highly of his name until I went to the seminar. A coworker who went a few months before me told me it wasn't worth the time, but it was the cult of personality that caused me to go. He ended up pretty much being a douche in person. Not one visualization we saw was created by him, he is literally curating other's great works, and for that, you can pick up used books on your own or visit a library.

7) All in all it felt way too touchy-feely and not very concrete. He only _briefly_ made mention of R and D3, saying they were high end, advanced visualization tools, that we as an audience shouldn't be using, because they are too complex. WTF.

8) Things that are NOT covered: a) Moving data, aside from his video visualization of music that he found somewhere. b) Visualizaing dynamic graphs or Sankeys. c) Interactive displays, such as on a website for users to use. The closest he came was to extol the virtues of ESPN. He literally recommended to bring up your presentations to the level of ESPN.com, and you'd meet with success.

TLDR; To each his own, but I wish I had my day back.


His books are great, mainly as a collection of interesting examples, but I agree with you on his seminar. I was let down. Nothing that I would really complain about since I didn't have to pay for it and it was really a day off work for me, but his seminar itself is not something I would pay money for.

I do think ESPN, weather.com, etc., are interesting examples because they have so much information to present, and it is nice that he was so forthcoming with concrete examples of what he holds up as great presentations of data. But I didn't find his analysis of them all that deep, and I think he made the fallacy of assuming that these sites are popular because of their presentation rather than, well, they're just websites of immensely well-known organizations. And weather.com is weather.com for God's sake--the URL alone guarantees readership.


I had weather.gov written down in my notes, I think we also saw weather.com. Thanks for clarifying. I also went for 'free' but that was 8 hours I could have spent writing code or playing with D3.

Did he also spend ~40 minutes ooh-ing over Google Image search? I remember seeing a bunch of blue sailboats or something and he was going on and on saying he prints out pages of google image search collages (except he had some artsy name that was more hipster than collage).


For anyone wishing to emulate Tufte's style in books using LaTeX, http://www.ctan.org/pkg/tufte-latex.


If you're looking for a more quantitative version of Tufte, I highly recommend Cleveland (http://www.amazon.com/Visualizing-Data-William-S-Cleveland/d...) and Wilkinson (http://www.amazon.com/The-Grammar-Graphics-Statistics-Comput...)


The difference between Cleveland and Tufte is that the former is "empirical" (he discusses lots of results from psychological studies on comprehension of graphics) rather than one man's opinion. Also, Cleveland actually has accomplished something other than scolding and praising.

Personally, I liked Cleveland's "Elements of Graphing Data" (http://www.amazon.com/Elements-Graphing-Data-William-Clevela...) better than "Visualizing Data", though they overlap a lot.


He's a legend from days of yore, but his books and presentations really don't talk about the web or even modern technology. It's great if you're working in traditional graphic design, and there are some core concepts you can take and apply to web design, but I don't see why people think he's some hero of modern design.


I found Designing with Data (http://www.designingwithdata.com/) to be a simple and practical introduction to data visualization.

I attended Tufte's seminar years ago, and found it entertaining, but that's about it.


Another good source on visual presentation (of data) is Cleveland (author of "The Elements of Graphing Data", see http://www.win-vector.com/blog/2013/02/revisiting-clevelands... ). The neat thing is Cleveland formally tested a lot of his visualization principles (asking people to estimate things from different graphs).


I've talked to multiple people who have been to Tufte's seminars. They all, unfortunately, say the same thing: in person, he's not a very dynamic teacher - maybe even a bit boring.

That does not detract from his ability to bring data visualization to the mainstream.


Applications are open for YC Winter 2016

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact