Updated June 6, 2019, mostly for typos. Apologies for the inevitably remaining typos.
When I went to find recommendations for book resources myself, I found primarily opinions and uncalibrated numerical scores of satisfaction offered in isolation. Occasionally these isolated reviews, and more often the table of contents preview, offered a little insight into what an author excelled at. Overall however I felt unprepared to answer one of the most common questions I am asked by colleagues or audiences: 'where should I learn more?'
My attempt here is to provide a relative perspective of what each of these resources excels at. I've tried to cover most of the resources I found commonly recommended by data viz twitter and the book buying algorithms. If I'm missing an important book let me know on Twitter and I'll enqueue it. Following me there is also probably a good way to keep up with additions to this list.
I have been very intentional to include tooling agnostic guides here. Many of us are constrained in the tools that are available, or familiar to our collaborators. Furthermore, too much focus on the 'how' may distract from the important transferable skills of data presentation generally.
There are a lot of interesting online resources or individual articles which provide either a very shallow overview, or isolated deep dive. I have not found any yet that provide a consistent or comprehensive foundation in the way many of these books do.
Jump-start books are either pretty brief, or useful being flipped through as reference while you’re working, or both.
Foundational texts don't necessarily need to be read cover to cover for benefit, but provide a thorough framing, explanations of why practices are preferred, and importantly references to supporting or supplementary resources.
Specialty resources include both items which might be cited by the foundational ones, or are a tangential extension.
If you're brand new and don't already have opinions about some of these books to start to calibrate, my suggestion is to start with one foundational book at a minimum, and ideally also one of the Jump-starts.
The Guide is incredibly concise and formatted as dos and don'ts. Thin and portable, it also covers tables and some basic formulas and common pitfalls. Published by Wall Street Journal, so most examples and topics are business subject matter, but the guidance is quality and widely applicable. It achieves brevity by leaving out references, history, or the underlying cognitive science, but that makes it an even more efficient resource if you're leaving those for another time.
Good Charts' strength is in emphasizing the importance of context in determining what makes a good chart. It is explicitly open and flexible with many definitions, taking a very approachable and positive tone. Relatively prescriptive about process for design, such as how long to spend sketching and prototyping. Includes a scattering of references, case studies, interviews, and presentation tips, all in a somewhat awkward floppy landscape format.
Fundamentals presents a thorough coverage of standard and many scientifically common plot types. Fundamentals presents a helpful taxonomy of acceptable, ugly, bad, or wrong. Datasets are typically followed through several plotting iterations, rather than just a before/after. Although this review avoids tool focused resources, the production of all figures in R (source online as supplement, not distracting in text) and emphasis on reproducibility means the examples are all very practically achievable. Fundamentals doesn't make it to the list of foundational resources because it is light on references, history, or the underlying cognitive science, though it does include a short bibliography. Occasionally veers into equations assuming moderate mathematical familiarity, but overall presentation is very approachable.
A handbook, a textbook, and a coffee table book had a perfect baby. It even includes a bookmark, because the authors apparently love their readers. It is simply hard to imagine a more densely informative, thorough, practical, and pleasurable resource. Downside is that as of now must be ordered from Finland , which in my opinion is worth the $30 for several week shipping. Acording to the authors it will be available stateside soon.
Approaching visualization as a journalist, the writing in Cairo’s books (there's another one on the list) is some of the most pleasurable to read. He intends information graphics and visualizations not only to illustrate a predetermined story, but facilitate exploration and discovery in what they present. Furthermore he presents the narratives of his own explorations. In these ways, the books remain relevant to exploratory data viz. Both books contain lots of citations for those looking for a jumping off point to further resources. The Truthful Art covers not only visualization, but some data summary and statistic foundations.
Stephen Few provides actionable guides that are reminiscent of Tufte’s work in a tendency towards minimalism and a tendency towards novel representations. Direct Tufte canon references are included, thought there are also some explicit counterpoints and Few is generally well grounded in the evidence based practices. Each book includes helpful page margin notes and citations, some rather dated of poor-practice examples.
Show Me The Numbers is a general guide to data visualization, though most of the advice is phrased a little more towards explanatory than exploratory. The format and writing of the book has a helpful and conversational tone. It includes quiz like formats for reinforcement in some sections, and concise chapter summaries throughout. There is a refresher on most common statistics, plus emphasis also on table design and when to use them. Presentations such as vocabulary-to-quantitative-plots-type associations may be particularly valuable for audiences who don't already regularly converse about data.
Now You See It focuses on exploratory data analysis, and provides a somewhat condensed set of basics compared to those in Show Me The Numbers. This book is particularly valuable in its reliance mostly on familiar plot types, which are likely to be available to users in any modern software of language. In so far as it presents refined ways to use common visuals it is a far more practical dashboard guide than Few's book explicitly on dashboards (which you'll find far below).
A phrasing of the data visualization more like a design process. This is not to imply it focuses more on the aesthetics, but that it prompts and guides very explicit consideration of the audience motivation and context for data visualization work. Kirk takes a compromising and realistic approach, as well as providing additional online resources (not reviewed). The book is a little bullet-pointy, but at the service of summarizing frequently. A substantial section covers what Kirk calls the 'Hidden Thinking' around goals, audience, and data handling, and editorial sense. There is a plot type quick-reference in the center of the book for, and a section devoted to data literacy as a consumer of data visualizations. Because the overall the focus is weighted towards the reasoning around data design, and the plot examples used are mostly from external publication sources, this book is not as much a source for consistent style inspiration or rapidly applicable dos and don'ts. Colorful quotes from well established sources such as authors on this review dot the book, but doesn't cite out to further reading directly in text (it's also online).
Makeover Monday grew out of weekly reworking of a broad range of data visualizations. Those 'Makovers' were presented online and inspired a participating community. This book is a more introductory level than most of the others in the Specialty group, but it didn't quite fit the developed thoroughness of coverage that characterize either the Jump-start or Foundational materials. Despite a bit of awkwardness as to what level it addresses, it has some very valuable aspects. It contains helpful emphasis on how to approach a new dataset, high level quality or 'sense' checking data, and iteration. Another particularly helpful aspect is that many of the makeovers presented improve a visualization without completely restyling it. Because of its origin, it is also an introduction to a distributed data viz community as a potential space to grow your skills outside of your immediate professional environment and pressures.
Andrews presents a historic, linguistic, and cultural context for how we communicate information. The book leans towards, but does not exclusive focus on, the quantitative. It is infused extensively with erudite references and quotations, which are enjoyable but also a bit distracting at times. There is good advice on specifics, though it's pretty condensed and often given abstractly or generally. Particularly noteworthy are the charming had styled illustrations throughout, including a short section at the end on how they were produced. The drawings make this book especially timely given the recent popularity of sketch-noting.
Per the title, the focus here is on Storytelling, and many of the specifics are for presentations. This book starts from the point where you have quality data that you understand to tell a story. You'll find concise but quality treatment of visual good practices, and mentions of other notable presentation resources. While there is emphasis on the importance of context, the advice that it should be largely omitted from the final story seems of questionably general. I would suggest this book those already well versed in exploratory data analysis and quantitative fundamentals, who now need to expose the few gems compellingly to an audience several steps removed.
Munzer provides a textbook with a rigorous formal structure for the discipline of creating and evaluating visualizations. There is an emphasis and greater coverage on interactive and scientific exploratory applications. Academic feeling design sensibilities mean that while it's a thorough and precise treatment of the subject technically, it is not a source for inviting visual inspiration.
As with his more general book, Cairo’s writing is a pleasure to read. Treating explanatory visualizations as also intended to be explored, it is more tolerant of decoration while also explicitly expressing respect for the intellect and curiosity of the audience. The Functional Art covers design principles and practices grounded in cognition, with the context of larger compositions sourced from popular publication infographics and visualization. Midway through, the book transitions into case studies and interviews.
This text focuses on the human perception science. Expect lots of diagrams of the human visual system, and some optical illusions especially in the early sections. As the title indicates the coverage information representation generally, which encompasses data presentation and visualization. Ware does call out advanced topic practical tips for specific data situations in highlighted boxes throughout the text. For those interested, Ware preliminarily addresses topics of interactivity and emerging technologies like AR from a human information perception standpoint.
An interdisciplinary collection of articles, from philosophy to cognitive science to engineering and more. What one could easily pass over expecting to be terribly dry is not, though it is text heavy and the few example figures are poorly printed. The first article Visualizing Thought provides a more general context than nearly any other book on this list (Andrews covers a similar scope if you're not after the rest of the articles in this compilation). The rest of the collection contains deep dives that range from the analysis of eye tracking data, to the specifics of algorithm visualization in textbooks. One gem treats the often flummoxing subject of ‘Viewing Abstract Data as Maps’ complete with examples from XKCD.
Tufte is probably the most often mentioned name in modern data visualization, though at current his books (and course) are not representative of evidence supported best practice, and they have never been particularly practical manuals. This statement may surprise or offend some. The point is not to demean his contributions, rather to offer guidance I wish I had received. Think of how economist Adam Smith pioneered important concepts, and respecting this contribution does not mean The Wealth of Nations is an introductory or particularly current resource.
Each of Tufte’s books are full of historical examples, fine art reproduced for analogy or analysis, and eclectic collections of quotations. They are enjoyable to read, and lead one to feel marvelously cultured. Each is an exceptional aesthetic work. Tufte’s data visualization explorations tend towards either extreme sparsity or density of information, and in both cases toward novelty.
After a lengthy introduction of examples, this book introduces and explores Tufte’s dogma of the data-to-ink ratio. There is little to no treatment of human perception, color is not addressed, and even greyscale is infrequently employed. Plotting and figure suggestions rather extreme and tend to ignore the human cognitive burden of deviating from familiar plots for rather small gains in data-ink efficiency.
Here Tufte introduces his other lasting conceptual contribution, small multiple ‘sparklines’. This book gets slightly more evidence based, and contains a good section on corruption of evidence and arguments. It focuses intensely around case studies of communication failure, and extensively treats Tufte’s well known disdain for Power Point (slide based presentation more generally). While it is a tool that is often misused, he offers little actionable constructive guidance and equates the technology directly with Fascism.
I should reiterate here my bias in this set of reviews for books that provide or support actionable guidance on presenting quantitative information. The fact that I wouldn't typically recommend these doesn't mean they're harmful or even unenjoyable. I figure that (as with data visualization) context is helpful, so it's more useful to see the full set of what I've considered for context, rather than solely the ones I encourage others to seek out.
The introductory examples are distractingly dated, which detracts from their relevance as a teaching tool. The basics revisit a lot of what's in ‘Show Me the Numbers’ but don't cover it quite as thoroughly. In this book too Few frequently references Tufte and suggests high info to ink, then similarly prefers very sparse or dense designs which are clever but are not standard. The book is more likely to be useful if you have total design control, but isn't very applicable in thinking about how to proceed in a constrained set of options. It is worth including here mostly as a caution in that I purchased it enthusiastically based on its focus but would find Few's 'Now You See It' of much more use in any dash-boarding context I have encountered or can imagine.
Tufte concedes that color might be ok in diagrams, but only sometimes. Academically self referential but also has pop-ups. I love Lichtenstein more than most people, but if what you're looking to do is use data in your job, the explanation of art self and historical references in some of his commissioned lobby paintings is a rather wide detour... This isn't to say I don't enjoy perusing this visually pretty collection, perhaps for some very high level absorption of inspiration, but it's not a directed way to learn about the quantitative or evidence grounded best practice.
Another book with pop-ups from a man who rails against gimmicks. It contains a chapter called visual and statistical thinking (in case you read the table of contents online), which visits some of Tufte's well worn case studies, however does not provide the type of formula basics you will find in many of the Foundational texts or even Jump-Starts.