Is Tufte still required reading in 2022?
June 21, 2022 11:04 AM   Subscribe

The Visual Display of Quantitative Information was a landmark book when it came out. Does it hold up in this world where graphing and charting have become so ubiquitous ? I have a young friend who is taking an intro stats/data visualization class. I'm wondering whether I should buy them a copy of this book to set them on the path of righteousness. Is it too dated? Is there some other landmark book in the field that has taken its place?
posted by Winnie the Proust to Science & Nature (21 answers total) 51 users marked this as a favorite
Bertin's Semiology of Graphics could be another book to consider. It's somewhat deeper and more theoretical than Tufte's stuff.
posted by panic at 11:36 AM on June 21 [3 favorites]

I think Tufte still holds up, this is thoughtful and I think a good idea!
posted by sixswitch at 11:37 AM on June 21 [4 favorites]

At the time it was published, Philip Morrison — member of the Manhattan Project and longtime book reviewer for Scientific American — strongly criticized Tufte's book for its insistence that illustrations should be seamlessly integrated into text, and that even so much as a caption was verboten. He had other reservations too, as I recall, but I don’t remember what they were.
posted by jamjam at 11:53 AM on June 21

They're good to read. Round out with some Alberto Cairo, Stephen Few, Andy Kirk, or Scott Murray.

Some of the ones I've mentioned have (academic) beef with Tufte. I think some find Tufte curmudgeonly and overly rigid. However ubiquitous graphing and charting have become so, so many people still do it poorly. So, imo, the more Tufte/Cairo/Kirk exposure, the better.
posted by tayknight at 12:14 PM on June 21 [10 favorites]

imo still great, though like anything canon should be questioned as needed. The foundations are still very valid, and the quality of the books definitely helps sell the idea that quantitative graphics can be great.
posted by wemayfreeze at 12:19 PM on June 21 [2 favorites]

It (and the later books) are expected on the shelves of academic scientists who care about such things. I assign readings from it to young people and they seem to like it and enjoy criticizing it. But, I'm old.
posted by eotvos at 12:26 PM on June 21 [9 favorites]

This is a bit like asking "when everyone has Excel, do we still need to read How to Lie with Statistics?"

The ready access to charting tools makes it more relevant, if anything. Just peruse WTF Data Visualizations if you need a reminder. I've seen smart people who I respect make bewilderingly bad graphs.

I do think some of his hangups are a product of the time, like his emphasis on not wasting ink.
posted by adamrice at 1:01 PM on June 21 [9 favorites]

You might take a look at Tufte's follow-on books, too: Visual Explanations and Envisioning Information.
posted by jquinby at 1:18 PM on June 21 [4 favorites]

I do data visualization for a living, and have seen both presentations and books that referenced Tufte just this year. Still a classic, I think.
posted by missrachael at 1:26 PM on June 21 [4 favorites]

I think Visual Display, Visual Explanations, and Envisioning Information all still hold up and contain a lot of great stuff that I don't think will ever not be useful. (I can't remember what Beautiful Evidence (the next in the series) is about but probably just as useful.

I wish all the people I work with who use Powerpoint and Excel would read them and follow at least some of the principles.
posted by jonathanhughes at 1:27 PM on June 21 [3 favorites]

I went to a sold-out Tufte seminar in 2018; I think you're still good! I regret that I gave away my books when I moved abroad. Even though I don't agree with all of his points, it's a helpful framework and a good introduction to a lot of principles that people may not otherwise present to you.
posted by quadrilaterals at 1:33 PM on June 21 [2 favorites]

Echoing— Even more necessary in the age of ubiquitous infographics. I disagree that principles like “use less ink” are dated, the theory behind it holds.

Grammar of Graphics is another good one, especially if using R, where ggplot2 is de rigeur.
posted by supercres at 1:56 PM on June 21 [4 favorites]

He might want to consider doing the online version of Tufte's class, which comes with all of his books, including a recent one that will definitely be up to date.
posted by Candleman at 3:14 PM on June 21 [2 favorites]

I haven't revisited my Tufte books in a while, as they're a little too far abstracted from the questions I usually have now, like "how do I make this plot in R much cleaner"? Claus Wilke has a nice book (Fundamentals of Data Visualization) that is available in print and free online that I reference much more often now, though I'm sure The Visual Display of Quantitative Information still shapes my thinking about figures.
posted by deludingmyself at 4:58 PM on June 21 [6 favorites]

I do data analysis and create charts, diagrams and info graphics. I've been doing this in universities for over a decade. I've read Tufte's work and there's a bunch I don't like about it, but to come to that conclusion, I had to read and understand what he was on about. That was valuable. So even if the recipient doesn't like Tufte, if they are open-minded, any of his works would be a valuable gift.
posted by b33j at 3:10 AM on June 22 [1 favorite]

I also do data analysis and make graphs for a living. I think it holds up and is worth reading, even if I don't agree with everything.
posted by quaking fajita at 6:50 AM on June 22

I'm just now revisiting Tufte after the initial publications of his first 2 books.

I am a little disappointed by the online course, as it seems to be just a recording of one of his past seminars - I wish he had taken the time to re-record and optimize it for online learning. Plus, the 4-hr billed time is not nearly as impressive given that periodic 20-min "study sessions" are built into the video.

Fortunately, the course is priced in a way where it is nearly a freebie if you buy all his books as a set. (Work paid for mine without question, because the price point was a full factor cheaper than the other training I had requested.)

That said, I echo everyone upthread. Great reading, gorgeous books, even if there are things I disagree with; there is a lot of value, especially if you regularly present data to non-data people, or have to convince folks to not farm out their data visualization to marketing agencies who prioritize "chart junk" to storytelling.
posted by Wossname at 9:14 AM on June 22

Mixed advice: I like Tufte's books, although I still haven't bought Beautiful Evidence. I already had copies of two books when I went to one of his one day courses (and thus acquired all three then in print). I don't know if the course still has the same lesson plan, but the set piece on the Challenger disaster was convincing even though I'd already read largely the same thing in print. I did find the in-person course a bit too self-promotional and congratulatory for something that cost money to attend. A not insignificant portion of the day was dedicated to selling his expertise and patting the audience on the back for recognizing it, and that was a little off-putting.

But as people point out, attending the course and getting copies of the books that way costs essentially the same thing as just buying the books outright. I guess if you know going in that there's some self-promotion involved, and you're paying for the books either way, then going course-first instead of books-first might be the way to go.
posted by fedward at 9:26 AM on June 22 [3 favorites]

Having attended Tufte's course back in the day (~2005ish?), I think the general principles still hold. However, I have lately been getting a lot of practical value from Stephanie Evergreen's various resources, including her newsletter and blog.
posted by acridrabbit at 10:52 AM on June 22 [2 favorites]

I’ll second acridrabbit. I went to a Tufte training in 2015, and find Evergreens info to be a strong and more modern approach to data visualization these days. I’ve been pointing my staff towards a number of her trainings and resources lately. There are also great communities specific to different platforms - I find myself mixing with a lot of Tableau folks on LinkedIn and Twitter but they touch on general viz concepts and Tufte-style ideas frequently.
posted by JannaK at 7:24 PM on June 22 [2 favorites]

Leland Wilkinson's The Grammar of Graphics has influenced visualisation software libraries and provides a worthwhile perspective on how information is perceived and displayed. I got it via interlibrary loan and found it to be good for clarifying how to achieve things with modern software (as a background in theory rather than a how-to for any particular system).
posted by Wrinkled Stumpskin at 1:20 PM on June 23 [1 favorite]

« Older Online Project Tracking App Recommendations   |   Visiting Los Angeles with teen Newer »

You are not logged in, either login or create an account to post comments