How Useful Is Tufte for Making Maps? (2007) | Hacker News Reader

How Useful Is Tufte for Making Maps? (2007)

makingmaps.net

110 pointsfloatingatoll36 comments

For anyone who hasn't had the opportunity: I highly recommend taking Tufte's one-day class. You get a copy of his books and get to see at least one or two truly historic texts (it's been 20 years for me, so I don't remember off-hand what he had with him).

Plus he's a tremendously bright man with entertaining anecdotes.

So, even though I don't think that anything he shared that day has been useful in my career, it was a life-enriching experience.

Interesting; I've also been highly discouraged from attending his class from a recent attendee. She felt he wasn't invested in actually presenting the material and it just felt like an advertising racket to buy his books. Things may have changed in the past 20 years.

I attended the "See, Think, Design, Produce" one-day session that Tufte put on with guest speakers Jonathan Corum, Bret Victor, and Mike Bostock (back in 2014). It was a great set of lectures, but TBH, felt that Tufte's talk was the least engaging/practical/applicable.

I'm definitely a bigger fan of Tufte's books than his in-person lectures.

It would be a shame if true. I suppose if I gave the same talk for 20+ years my enthusiasm might wane.

I went to his one day event maybe 6 years ago, with a group from grad school.

Had the same experience, while it was cool to get all of his books, the actual presentation really wasn't that great.

There was also like 200-250 people in the room at least. Way too big for "intimate" learning.

I went earlier this year and I thought it was a good experience. The key is to arrive on time and sit there and do the silent reading that he assigns before the lecture starts.

I sat in the front and he approached me. He was friendly, asked me what I did, made some suggestions on books and articles to explore. He also signed my book, which was cool.

I'd take the class again for sure.

I think it's important to just go in with an open mind and not treat everything he says as gospel. I think he presents some ideas that very much cut against the conventional wisdom in Silicon Valley and that was nice to hear.

So yeah, I think it'd be worth your time to go. Plus, just because someone else didn't like it doesn't mean you won't get a lot out of it. Unless you go and see for yourself, you can't really know.

Fwiw the books are included if you buy a ticket, and the books are the medium of much of his work so it would make sense for him to extensively reference them

I would agree with that assessment. When I went, I attended in a huge room with hundreds of other people. It was not a seminar in the sense of being able to interact with the teacher, but a lecture. And the lecture did not present a sequenced set of actionable principles, but bits and pieces consisting of the history of design, critiques of existing designs, an occasional bit of advice, and a lot about Tufte. Overall, I did not feel that I'd received enough value to offset the entry fee and the time.

Edit: clarity

I took the class earlier this year. Like others have said, it’s a multi-hundred people affair, not very intimate.

Having read the books, the lecture provided no value whatsoever for me besides the “cool” factor of seeing ET in flesh (and a sneak peek at his next book). At best I would say that the lecture is a good crash course for people who haven’t read his books yet (the fact that you get the books as part of the fee supports that theory), at worst it’s mostly self promotion (he spent a solid 10 minutes showing us his twitter timeline and pointing out his favorite tweets that he had retweeted o_O).

I would be more interested in the course he puts together with other designers, if he ever does that again.

On the other hand, I found his one-day lecture banal and wildly over-priced.

Same here - loved the book and was very excited for the one day lecture but it was probably the most boring and banal one I’ve ever attended.

Agree. More like some kind of a monotonic presentation rather an educational one. Books are great for reference though. If you had a choice, just buy all the books and skip one day session.

Tufte is focused on his sculpture garden (http://www.atlasobscura.com/places/hogpen-hill-farms). At an event I attended he got frustrated with someone asking questions about data visualization, replying, "That's my old life, I'm an artist now."

I'm a bit more mixed. I found his presentation style pretty terrible and his specific contributions (viz and software) not terribly exciting. But he's a master at analyzing other people's visualization work with an almost mind-altering critical eye. His books are incredibly valuable.

I think it's probably just worth it to buy the books and read them, but I did find his walkthrough of a few visualizations enlightening.

I still recommend the class, but with the caveats mentioned above.

I feel this is true of Tufte in general (and I own his books and have taken his seminar): he has made, and finds, excellent examples of good design but his heuristics (they are never as strong as "rules") are rarely actionable.

I've had great success applying his principles to conditional formatting in spreadsheets, unexpectedly.

If you want actionable, I highly recommend Jean-Luc Doumont's "Trees Maps and Theorems." It's like a way more practical, actionable version of Tufte's work. Probably my favorite book on communicating with data. He's a great speaker as well, do take the opportunity to see one of his talks if you can (I believe some are on YouTube).

I completely agree! Doumont often gives talks at universities that are open to the public.

http://www.principiae.be/X0200.php

I really don't get people's admiration of "The Visual Display of Quantitative Information." I bought and read it years ago with great excitement, but felt very let down by it. I wrote a detailed Amazon review about it: https://www.amazon.com/gp/customer-reviews/R11NYC3OE3LBE/ref...

I went to a Tufte talk and it was much better. It included "free" copies of his books and to be honest I wasn't very impressed with them either. They didn't convey his main points very well. The biggest takeaway I got from his talk was that it's better to substitute a stack of low-information powerpoint slides with a few high density images along with a paper written in normal English. But to be honest, it's way less work to make a stack of low-information powerpoint slides, so that's here to stay.

I think you basically entirely missed the point of the book, to the point of obtuse deliberate misreading. Given the amount of work you clearly put into your review, and your apparent ability to write (and presumably comprehend) coherent sentences, I’m baffled.

You don’t seem to understand at all what Tufte means by chartjunk (even though he prints dozens of explicit examples and explains what is wrong with each one at length), you don’t seem to understand what he is getting at with the thought-experiment-like demonstrations of chapter 6. You have an overly literal and pedantic reading of his tongue-in-cheek dry humor. And you seem to have completely skipped the detailed analysis and explanations of the principles laid out in the book if you think that “Tufte's principles totally ignore the primary purpose of graphs, which is to show a data set's patterns (or lack thereof) to humans.” My one-sentence summary of the book would be “information graphics need to spend more effort on showing patterns.” That’s the central abiding message of the book, which Tufte hits us all over the head with again and again, on almost every page.

> My one-sentence summary of the book would be “information graphics need to spend more effort on showing patterns.” That’s the central abiding message of the book, which Tufte hits us all over the head with again and again, on almost every page.

I just searched the book for the word "pattern" and it appears on only six pages in the entire book. Tufte throughout the book is far more impressed with how many data points you can jam into a graph than how clearly the graph conveys a pattern or idea.

I'm also curious which of Tufte's statements I quoted that you think were intended to be taken tongue-in-cheek. They all seem pretty straightforward and factual.

Your reply reinforces my belief that most people's admiration of this book has to do with what they imagine it says, rather than what it actually says.

I would reply at length and quote extensively from every section of the book, but 15 years of internet arguments has convinced me there’s really little point and it would be largely a waste of time. If you can even flip through the pages in a few seconds and come to the conclusion that Tufte doesn’t care about how clearly graphics convey data, there’s really nothing I’m going to say that can convince you. So I’ll just leave it at “bae, you cray-cray”.

I can only recommend re-reading the book closely and carefully, ideally spending 5 or 10 minutes really looking at and thinking about each graphic. Maybe start with the first few pages where Tufte lays out his criteria for graphical excellence.

You may also want to look at his 2 subsequent books Envisioning Information and Visual Explanations wherein he goes into considerably more detail about particular aspects and techniques statistical graphics can use to show various types of patterns.

I have probably spent 50 hours over the years studying Tufte’s four data-viz books, and every time I look I find something enlightening, fascinating, or inspiring. Some individual pages reward 30 minutes or more of careful study. If you can’t find anything useful inside, I don’t know what to tell you.

> If you can even flip through the pages in a few seconds and come to the conclusion that Tufte doesn’t care about how clearly graphics convey data

That's not what I said. If you project your own meaning onto my 100-word comment, it does not surprise me that you project your own meaning onto an entire book.

----

I'm glad that you like the book. If your goal was to share your positive feelings about it, "obtuse deliberate misreading" and "bae, you cray-cray" may not have been your best strategies. If you want people to be kind and understanding, don't start with insults and insinuations.

It is also not convincing when you say that I missed the tongue-in-cheek dry humor but decline to say which of my quoted passages you think were intended to be humorous.

I read this book because I had high hopes for it. It disappointed me a lot.

I understood your sentence. You didn't understand mine: I made a clear distinction between data and patterns/ideas. I think Tufte overemphasizes data and underemphasizes patterns/ideas.

Colleagues of mine treat Tufte's work like a literal ten commandments. I can't really fault his logic on the minimalism and he's certainly earned his rep in the GIS space, which really can be applied to pretty much any visualization.

I'm just curious why this post is making a recurrence now, is this because of the twilio post that made its way to the HN frontpage?

I hadn't seen that, thanks! For others, it's at: https://news.ycombinator.com/item?id=15274509

> Graphical excellence is that which gives to the viewer the greatest number of ideas in the shortest time with the least ink in the smallest space.

I think that's good for web design too.

"Commandment 5: Map Layout Matters, includes (19). Layout is a bigger issue than this one point from Tufte and it is an issue that is not stressed in map design texts (although I do devote a decent chunk of a chapter in Making Maps to map layout). It is difficult to talk about map layout (or the layout of information graphics in general) in the abstract. But layout strongly effects the look and feel of the map, and can make a map easy or difficult to read and interpret."

Most (all?) states in the US provide free road maps at the welcome centers on major roads entering the state. For many years, the map of the state of Texas had the panhandle, the squarish bit that sticks up to the north, separated and plopped in the middle of where New Mexico should be.

Somebody apparently complained and then the new maps that came out contained the whole of Texas, well connected. As a result, these maps were much, much larger than the previous maps, given that you're kind of limited in the scale of a road map and that they were now including a majority of the states of New Mexico and Oklahoma.

Sometimes the best presentation isn't the best presentation.

A more useful book is Alberto Cairo's Storytelling with Data, great examples and the reasoning behind it.

And yes I have Tufte's books too.

Tufte is a collector of some good design, but having seen him at a conference and gone through his books, he actually has very little grasp of the general principles at play. I wouldn't recommend him.

That seems unlikely, but I'm curious. Do you have an example of a general principle that you think he hasn't grasped and why?

OT: Has anyone done a similar distillation for the 12 factor app? As I was reading up on 12 factor, it seemed like you could distill it into 4 principles or so.

Can you expand on those 4? 12-factor has been a great guiding principle for me the last few years.

I have definitely done that in my mind. I don't agree with all of the items in 12 factor. The principles that drive those 12 factors are ones that I follow, just not the specific implementation that is presented. I don't know that it distills down to 4 or whatever number, but I'm certain it is less than 12. For example I don't think that it's always appropriate to store config in the environment variables. But the principle of not embedding config in code is one that I follow.

Reply on news.ycombinator.com