Edward R. TufteCritiques

How the mighty are falling... so many typos; did he ever proofread this? And too many pages (of this admittedly beautiful book) I found to be incomprehensible. And of those I did understand, many seemed to be simply conclusory material: too often he presents statements or quotations as if they are facts, or clever-but-true aphorisms (which are something else entirely), without providing any basis for them.

His first volume is amazing. The second is good, although the fourth is mostly a somewhat improved version of similar material. The third left no impression on me at all. And this fifth seems to be written to a much different, and far lower, standard. Much of the time, I simply had no idea of what new point he was trying to make (nor, therefore, how it might affect how I create or consume material intended to communicate facts).

I try to read Tufte's books with my brain firmly engaged; but even so this one just left me wondering what it was all about.

The fourth of Tufte's series of books on the visual display of information. It is the most eccentric of the series and includes a moderate amount of recycled material from the previous volumes. There is remarkable vitriol in an amount that I think you can only see in a self-published book, with an illustrated non-anonymized attack on one economics professor's book, and a prolonged attack on Microsoft Powerpoint that refers to Stalin more than once. There are also, of course, many interesting things, e.g. some discussion of Conway's Law. The book ends with a very odd criticism of some works of landscape architecture, and then photographs of some of the author's sculptures.

The beginning textbook for any UX designer or anyone passionate about how information and patterns can be structured.

During years of combing thru data visualization books, I had seen this book mentioned various times. Finally found a used, hardbound edition in excellent quality. I look forward to paging thru the book in detail during a break.

I feel so validated by reading this book. For years, I thought it was just me that I could not get past the superficial way data was portrayed on Powerpoint (PP). I could not find any way to aptly analyze data presented in that format. This book helped me to see that it wasn't me, but the style of the data presentation which is so limited by PP. Tufte states that, "PP templates for statistical graphs and data tables are hopeless." He goes on to explain that PP, since it is proprietary, has no incentive for meaningful change, especially since it is essentially a monopoly.
He uses the example of NASA's reliance on Boeing's examination of the Challenger disaster using PP rather than technical papers, which unfortunately underscored the danger of the situation encountered by the Challenger team.
The information in this book will drastically change my practice of using and being a recipient of PP presentation from this time onward.

A wonderful classic; by now rendered mostly redundant if you have read recent literature or work on visualizing information (granted, it was inspired by this). It was useful in providing me with several axioms to consider when attempting to visualize data... not because I necessarily agree with a specific approach.

This brief book about specific case offers little in the way of grand theories and too few examples for readers to deduce their own.

A fascinating, elegant, and beautiful book.

Picture books for grownups! I'm reading another one of his books right away, this one was so fun. Well, it's fun if you enjoy charts, and learning about things like who invented the bar graph, and looking askance at poorly-made graphics.

This is Classic in the field of Visualization. I highly recommend anyone into UX/UI to go through this book. It's simple read and won't take much time.

Outline:

Part I — Graphical Practice:

1 Graphical Excellence
2 Graphical Integrity
3 Sources of Graphical Integrity and Sophistication

Part II — Theory of Data Graphics:

1 - Data Ink and Graphical Redesign
2 – Chart junk: Vibrations, Grids and Ducks
3 - Data Ink Maximization and Graphical Design
4 - Multifunctioning Graphical Elements
5 - Data Density and Small Multiples
6 - Aesthetics and Technique in Data Graphical Design

Part III — Design for Display of Information
1 - Epilogue



Part I — 1) Graphical Practice: Graphical Excellence:

Tufte’s summarizes by saying, it is a matter of substance, statistics, and of design. It gives the viewer the greatest number of ideas in the shortest time with least amount of ink space. It consists of communicating complex ideas communicated with clarify, precision and efficiency.

- It is about clarity in communicating with precision, efficiency that shows data, induce user to think about substance of visualization, avoid distortion, present many numbers in small space, make large datasets coherent, encourage eye to compare various pieces of data, reveal data at several layers of detail, serve clear purpose, be closely integrated with statistical and verbal description of data.

Graphics reveal about data.

-Every Graph or Visualization should let the user to think about the data, not the methodology or technique
-Time Series Data

Two Greatest Scientist of Modern Graphical Design are J.H Lambert, Swiss German Mathematician and William Playfair, Scottish political economist.
Playfair preferred Graphics.

-Descriptive Chronology is not casual expression.

Charles Minard, the French Engineer who explained Napoleon’s army — combination of data-map and time series.

Most Modern Graphics are relational — x, y that encourages to find out casual relationship.

Part 1 — 2) Graphical Practice: Graphical Integrity:

For many of us, we constantly think of lies when we look at a graphic or statistic.
Around 1960, John Turkey made Graphical Practice respectable.

What is distortion in Data Graphic?

Lie Factor = size of effect shown in graphic/ size of effect in data
If it’s greater than 1.05 and less than .95, then it’s substantial

Show data variation not design variation.

Context is important for Graphical Integrity — compared to what?

Lying Graphic cheapen graphical art everywhere.


The Six principles of Graphical Integrity:

- Representation of numbers present in the graphic should be directly proportional to numerical quantities reported
- Clear, detailed and thorough labeling should be used to defeat graphical distortions and ambiguity
- Show data variations, not design variations
- In time series display, use standardized monetary units
- The number of information carrying dimensions should not exceed number of dimensions in data
- Graphics must not quote data out of context



Part 1 —3) Graphical Practice: Sources of Graphical Integrity and Sophistication

Why do they lie?
-Lack of Quantitive skills, the doctrine that statistical data is boring

Many believe that graphics are there to entertain unsophisticated readers. Japan has the highest use of statistical graphics in their newspaper.

Part II — 1) Theory of Data Graphics: Data Ink and Graphical Redesign:

Data Graphics should draw viewers attention to substance of data. It should form quantitive contents.

Fundamental principle is, “Above all else, show the data.” This is the principle for a theory of data graphics.

Data Ink ratio is data ink/total ink used in graphic. Remember to maximize the data ink ratio devoted to the data. Other side of data ink ratio is to erase non-data-ink, within reason.



There’s five principles in theory of data graphics produce substantial changes in graphical design.
-Above all else show the data
-Maximize the data-ink ratio
-Erase non-data-ink
-Erase redundant data-ink
-Revise and edit


Part II — 1) Theory of Data Graphics: Chartjunk, Vibrations, Grids and Duck

Interior decoration of a graphic produces a lot of raw ink that does not tell the viewer anything new. The Grid might include a lot of chart-junk.

When Graphics are taken over design or styles rather than quantitative data, it is called as Big Duck.

Part II — 2) Theory of Data Graphics: Data Ink Maximization and Graphical Design.

Reducing ink ratio in some of the charts might induce changes.

Part II — 3) Theory of Data Graphics: Multifunctioning Graphical Elements

The Same Ink should be used for more than one graphical purpose, it carries data information and performs a design function usually left to non-data-ink. Data based grid is shrewd graphical devise.

Sometimes, the puzzle and hierarchy of multifunctioning graphical elements can data graphics into visual puzzles, crypto graphical mysteries. Colors sometimes generate graphical puzzles. The shades of gray gives us more easier comprehension. This is the key.

Part II — 4) Theory of Data Graphics: Data Density and Small Multiples

How many statistical graphics take advantage of ability to detect large amounts of information in small space? Let’s begin with empirical measure of graphical performance and data density.

Data density of graphic = number of entries in data matrix / area of data graphic

More information is better than less information. Maximize data density and size of data matrix within reason. High volume data must be designed with care. The cost of chart junk, non-data-ink, and redundant data-ink is even more costly in data rich design. We apply shrink principle, which means graphics can be shrunk way down. Bertin’s crisp and elegant line displays small scale graphics in a single page.

Small Multiples, resemble frame of movies, series of graphs, series combination of variables.



A Well designed Small Multiple will contain:

- inevitability comparative
- deftly multivariate
- shrunken high density graphics
- based on large matrix
- drawn exclusively with data ink
- efficient in interpretation
- often narrative in content.



Part III — Theory of Data Graphics: 1) Aesthetic and Technique in Data Graphical Design:

-Graphical Elegance is often found in simplicity of design and complexity of data

Visually attractive graphics gather power from content. Basic structure for showing data are sentence, table and graphic. Often two or three of this should be combined. Make Complexity accessible.

Graphics should prefer towards horizontal, greater in length than height.
Lines in Data Graphic should be thin. Graphical elements look better when their proportions are in balance.

I have all five of Tufte's books on graphic display and, while it is interesting, I find this one the least enjoyable. This is partly down to the scrappy layout and structure, presumably trying to emulate a notebook style of information presentation. At times it seems repetitive, and makes strange claims for disruptive typesetting.

A slightly dated text with very modern adaptations.

This (classic?) book focuses on what makes a good graphical display of data and has a lot of examples of how to do, including the mandatory horror examples of how to not to do.

I agree with almost everything he says though some of it doesn't really matter that much in computer generated graphics (saving ink to save drawing time). Unfortunately it becomes quite obvious that a lot of graphical display of data in blogs, newspapers or other media has been designed by people that never found this book.

If I'm to summarize his advice it's "draw less". Draw less to not confuse the reader with mixed signals and keep it simple for the same reason. He has a number of examples of what can be removed from a "default" graph without losing any information and all and in the progress enhancing the visualization of the actual data.

You'll love the first Tufte book you read. If this is your second or third, you'll feel like he is repeating himself.

A really interesting book, still useful despite being dated. A lot of Tufte's lessons have been internalized by contemporary designers, though there are certainly lots of terrible charts out there. Still, the use of moiré he inveighs against is thankfully long since abandoned. The book is not as useful today as it would have been to read 30 years ago, but still informative.

An absolutely essential manual for anyone working in graphics who wants to use data analysis or quantitative information in any way. The book doesn't offer concepts, just information, and it is clear and comprehensive.

Another very readable, visual, and to-be-absorbed book by Tufte on how to make effective visual displays of information.

All of Tufte's work should be required reading in high school.

A classic.
A beautiful piece of work.
A minimal primer.
A coherent aesthetic philosophy.
An amulet of protection against data-imbued confusions presented as "infoviz"es.
A 'start here'.
The last word.

If you are ever going to hold a presentation with slideshow software (like PP or Keynote), you must read this essay!
Its the best analysis of what this kind of software acutally do with the massage ever written.

This is a much better book than the first one, there's a lot more substance and less polemic. However, it's not very well written and his main points get lost in the long series of examples, so it takes some work to get that substance out of it.

I got a lot less out of this than I did from The Visual Display of Quantitative Information, which is to say still a fair amount. I thought it was an enjoyable read nonetheless and certainly worth spending a few evenings on. The main thrust of the book seemed to be that it's important to show the reader information in a way that maximally enhances their understanding. Beyond that there's a list of examples of what to do and not do, and I'll remember those down the line when I'm looking at or making pictures.

This was a much more engaging read than I expected it to be. It's not just a cranky old academic complaining about style. He really rips PowerPoint apart. The in-depth analysis of the NASA incident is especially damning. PowerPoints were used in place of technical reports when they were assessing the damage to the Space Shuttle Columbia's wing. Although the evidence did not truly suggest the shuttle would be fine, the takeaway from reading the PowerPoints was that everything would be OK. Instead the shuttle overheated and blew up upon reentry.

This was especially interesting for me, having just recently finished "Understanding Media" by Marshall McLuhan. I could sense McLuhan's ideas underneath Tufte's text. Tufte argues that PowerPoint is a marketer's tool for sales pitches, which are not intended to deliver true information. They exist to manipulate the audience. And that is what has become of our scientific, academic and professional meetings. We do not deliver evidence with PowerPoint, we deliver a sales pitch. The result is poor decision making. I'd love to hear both authors thoughts about Twitter (if McLuhan was still alive). This has certainly changed my perception of PowerPoint.

A book that rewards close reading.

Great read for anyone interested in data visualization.

Before you make that presentation, read this.