Storytelling with Data: A Data Visualization Guide for Business Professionals Paperback – November 2, 2015

People wanting a basic introduction to presentation graphics would be
well served by this book by Cole Nussbaumer Knaflic. Prerequisites are
minimal: there is almost no mathematical content and no use of any but
the most elementary statistical methods. Knaflic's encouraging message
is that MS Excel and PowerPoint can be quite enough software for good
graphics, but you will need to go beyond the defaults and work at the
details.

Almost all the examples are of very small datasets already to hand with
two-way structure. 2 variables for 12 months and 5 products for 7 years
are typical sizes. In practice when analysing data it can be hard work
deciding what methods to use and reducing a mass of raw data to a
concise summary. These steps, sometimes most of a project, are here
assumed already done.

The subtitle flags a focus on "business professionals"; the content
tactfully implies junior people presenting with PowerPoint to
time-challenged bosses at brief meetings. Seemingly few write reports to
be read any more, or use any other presentation software.

Knaflic is excellent on the need to keep things simple. She has a good
eye and sound logic on what looks and works well and what does not.
Examples show how mediocre graphs can be improved by reducing clutter,
killing the key, better use of color, and similar standard tricks.
Horizontal bar charts are usually more readable than vertical, and pie
charts and a false third dimension are best avoided: these points have
been well made many times, yet do deserve forceful repetition. Various
kinds of bar and line charts are her main work-horses.

Sometimes the discussion seems a little contrived, as poor graphs are
set up to be shot down, but that's often what convinces. Readers should
be on the author's side as she encourages us towards effective and
tasteful graphics. Her combinations of blue for data deserving emphasis
and grey for data providing context -- or of blue and orange for groups
to contrast -- are good design patterns for experienced analysts as well
as outright beginners.

The closing chapters are more long-winded and repetitive, but do include
small gems. A splendid case study on avoiding spaghetti graphics (lots
of tangled lines) stands out, and the problem and the ideas deserved
more.

I always find it disappointing when datasets are fabricated or
sufficiently anonymous that they might as well be. People care most
about their own data which an author cannot provide, and confidentiality
constraints often bite, but real data examples are still generally
preferable to fake. Too many examples here are variants on Products A to
E or Features A to O. Unfortunately an outrageous example of a bar
chart from a well-known U.S. television news network (p.50) seems all
too real.

What's not here includes Cleveland dot charts, histograms and box plots
even among the staples of good introductory statistics courses, let
alone (say) use of logarithmic scale, always one of the first graphical
devices for many sciences. So if you want something with more
statistical bite or depth, you need to look elsewhere. Naomi Robbins'
excellent, no nonsense Creating more effective graphs
http://www.amazon.com/Creating-Effective-Graphs-Naomi-Robbins/dp/0985911123
would enable you to go further.

As in any first edition there are some small slips and exaggerated
claims. 40% is not a majority (p.5). There is confusion between number
and percentage on p.39. Any rule that "bar charts must have a zero
baseline" (p.52) is simplistic. It is quite correct that bar charts
should encode departures from some sensible reference level. (The
television network responsible should have paid attention.) But that
reference level could easily be some value not zero, such as parity
between men and women, or the mean of a variable, or 32 degrees
Fahrenheit to separate freezing and non-freezing temperatures. I
disagree that every dollar amount or percent should be labeled as such
(p.90); that is repetitive clutter such as Knaflic rightly deplores. Nor
is it an absolute principle that every axis needs a title. If the axis
labels are 2008 to 2015, no one should need "Year" to explain what is
happening. Far from being "extremely rare" (p.141), several exceptions
to that principle are included in this book!

A note on style: Inside a very useful book is an even more useful
shorter book struggling to get out. For my taste, the motivational
warm-ups and little anecdotes are often too spun-out or too trite. Good
graphics should be presented as illustrations within a good story: a key
point, but not one that required a long chapter with digressions on Red
Riding Hood or on Aristotle on drama, or advice from a junior high
school teacher. A tighter copy-editor would have signalled that
"leverage" (used as a verb about 70 times) was too much of a personal
favorite, while "de-emphasize" for "tone down", "utilize" for "use" and
"incredible" for things all too credible are among several other
repeated tics.

An easy solution is to skip and skim: if a book is on graphics, you can
always just look at the graphics. In this case, Knaflic has written a
worthwhile book that, small details aside, does well what it tries to
do.

I am a university professor who teaches biostatistics and I find this to be one of the best books that bridges the gap between analytics and presentation. There are some excellent books around that show visualization (e.g., The Wall Street Journal Guide to Information Graphics: The Dos and Don'ts of Presenting Data, Facts, and Figures or books by Few Information Dashboard Design: Displaying Data for At-a-Glance Monitoring &Show Me the Numbers: Designing Tables and Graphs to Enlighten or Cairo The Truthful Art: Data, Charts, and Maps for Communication) and there are good books on presentation (in particular I love Duarte's books Resonate: Present Visual Stories that Transform Audiences) but this book is unique in how well it blends the two topics. I have never seen such an excellent presentation on how to build a series of graphics. That is, with books by Few or Cairo you will know how to make *a* great graphic and with advice from Duarte, you can connect with your audience but with this book you will see how to build a series of interrelated graphics that highlight different parts of a dataset. Most of the examples are spun around business but the examples are easy to extend to any field.

While I think the author wrote this for people who do presentations in any quantitative field for a living, this book should be required reading for graduate students preparing to defend a dissertation or thesis.

This is far and away the BEST data vis book because it goes beyond the techniques for making visualizations to the ways in which you can connect with clients, co-researchers, and stakeholders. This is an essential book for effective reporting and dissemination of information. Truly exceptional.

Very rarely do I find a business book that so fully meets my needs and expectations...and then surpasses them.

As a full-time business analyst who was hoping for some practical guidance to improve his visualisations, I was delighted by how clearly the author delivered it with her easily understood concepts and step-by-step demonstrations.

Whether you are already acquainted with data analytics tools like Tableau but think your presentations look boring or you're an artistic designer who can't figure out how to use software to express your point, you need this book!

As a professional working in one-to-one marketing, I feel like too often we succumb to flashy "junk charts" or otherwise ineffective presentations of data. This book seeks to break this habit by teaching typical BI professionals about how to more effectively communicate using data.

The depth at which the author discusses each of the primary topics is appropriate for the majority of applications and the visuals presented are top notch. That praise aside, if you're looking to go (even) deeper about individual parts and pieces of charts and tables, I'll also recommend you pick up Stephen Few's "Show Me The Numbers," as he gives an incredibly comprehensive view on the micro-level of visualization design. In my opinion though, this book covers enough ground for the majority of readers and is a great way to "up your game" when it comes to presenting to a variety of audiences. I'm going to recommend it to my colleagues as a way to lift the quality of data visualization in my department as a whole.