To calculate the overall star rating and percentage breakdown by star, we do not use a simple average. Instead, our system considers things like how recent a review is and if the reviewer bought the item on Amazon. It also analyses reviews to verify trustworthiness.
I found this book to be an excellent resource for both educators and students. Wilke’s book provides you with a clear, straightforward and in-depth introduction to data visualisation, that can be used to learn about or to improve upon your current understanding of the subject. What I like about this book is that the organisation of the content is divided into logical chapters, which provides a structure that’s easy to reference back to. The Illustrations provided are great as they help in explaining the concepts in the text well. I also like how the content is primarily focused on theory, practice and the aesthetics of good chart design, instead of how to build or code each visualisation.
This is a practical guide to making data visualisations that convey information to a wide variety of audiences.
I liked the focus on the generated images, rather than getting bogged down with coding details. The code is available on github if you want it, but the place for a book like this is to discuss the concepts and purpose of data visualisation such as scale, colour and representation.
I feel this book will have a significant longevity - if you are reading this review in ten years time, this book will still be useful.
According to the author, “data visualization is part art and part science.” And, in my view, the author excels at both.
What I particularly love about the book is the author's classification of visualizations: ugly, bad, and wrong figures. Pure genius!
The book is probably most useful for Data Scientists using R, Python, Power BI, and Tableau.
A few personal highlights:
- Page 29: Besides from Tableau, I’ve rarely seen so much attention to color-coding continuous and categorical data. I love it. - Page 53: If you’re in Data Science or Machine Learning, you will appreciate the Titanic samples. - Page 61: The importance of not relying on default bins for visualizing a single distribution. - Page 96: Many don’t like pie charts. I love the author's example where a pie chart might have been superior to other visualizations. - Page 112: How visualizing nested proportions (I really don’t like the “sunburst”) can be improved. - Page 233: Common pitfalls in color use. - Page 267: EVERY FIGURE NEEDS A TITLE. Something I still do wrong. - Page 297: Why line drawings have been popular. Here, the author shows a strong historical understanding of visualizations. What more can you ask for? - Page 338: Something I have to remind myself constantly: “never assume your audience can rapidly process visual displays.” - Page 343: how to strike the balance between making a visualization memorable and clear.
The book is clearly among my top 10 Data Science books.
Wilke’s “Fundamentals of Data Visualization” fills an important niche: a manual for the professional data scientist who wants to convey the essence of data succinctly, accurately, and in an aesthetically pleasing way. Wilke shows that this is (perhaps surprisingly) difficult, and he does this by showing both good and bad examples of data visualization in order to make this point. The book really attempts to build up the reader’s intuition about what makes a good visualization chapter by chapter. By the end, the lessons learned have become so obvious that when going back to earlier chapters, a single glance at the figure is enough to remind you about what is good (or bad) in the example given. This is not a book that gives programming examples that tell the reader how to achieve a particular figure, and I think this is intentional. The emphasis is on understanding the principles of data visualization, not on supplying a hack for the next figure. (But for those who want to recreate a figure, Wilke provides the entire source code for the book on a web site). All in all, this is probably the best book on data visualization for the practicing scientist out there. The prose is clear and concise, the principles behind the choices he makes are clearly laid out, and the figures are clean and free of clutter. My only gripe is that the colors in the book appear to be somewhat flat and less intense than in the electronic version. Hopefully this can be remedied in the future.
As a working data scientist and recovering academic, I've been making data visualizations for a decade. I've read lots of books about data visualization, but there were few I found difficult to put down; this book has been an exception.
The principles in the book are solid and agree with my experience: following the guidelines in the book will help you craft more meaningful figures that resonant more deeply with your audience and allow you the freedom to move away from the production of criminally overwrought visualizations that are so common in academia and elsewhere.
In the process, Wilke gives very approachable, high-level forays into elementary statistics, PCA, and map projections that felt like welcome and complementary diversions. I would have liked to have read a bit more about how humans perceive different visualization types, plotting characters and the like and how empirical studies of human perception can inform our design of visualizations; this was sprinkled throughout the text but a chapter on it would have been welcome. But that's a minor point; the text is clear and the entire book is incredibly well thought-out.
The code for all the figures, should you want it, is online; if you are an R user (note that the book is language-agnostic), studying it will give you power to craft and tweak your figures to a degree to make even the Wickham's of the world jealous. Highly recommended.
Full disclosure: 3 years ago I became a colleague of Prof. Wilke at the University of Texas, and we have interacted occasionally since then although we are in different departments. I recently purchased this book because of interest in the topic and my respect for Prof. Wilke and have not been disappointed. I have taken the author's advice and have not read the book cover to cover but have been dipping into chapters of interest--but then I find myself hooked and reading further on. The author's straightforward presentations and clear delineation between good, bad, and matters of opinion are helpful and spot on. I am working on two scientific papers at the moment and found advice and discussions in the book that have improved figures in both papers--and I had previously prided myself on what I thought was a thoughtful approach to making figures. I left the book with my students, and now a regular feature of our weekly journal club is a Wilke-inspired critique of how the figures could have been better made to illustrate the main point. I strongly recommend this book to anyone whose work entails creating graphical illustrations to present complex data. It is full of thoughtful analysis, good advice, and great examples and will make your presentations and publications better.
I saw this book on a friend’s desk and had to get my own copy! I’ve read straight through a quarter of it already and it is so clear and straightforward that it makes all the ideas seem obvious. I had previously tried to absorb some works by Edward Tufte but found them overwhelming compared to this book. I’ve been an academic biologist for years and wish I’d had a book like this earlier in my career. The principles the author explains help me codify what I already intuitively sensed about various visual representations. It will be much easier now to design effective figures for my own work and to interpret at a glance the data of others! In fact I have already used a principle from these early chapters to improve a figure I’m working on for a manuscript. While extremely readable and appropriate for any layman, this book is a must for professionals representing data in any form, and in teaching graduate students how to communicate data. This book is so straight forward I would hope it could be used even in high schools where the basic important concepts would serve everybody entering this data-rich world!