Data Smart, by John W. Foreman (Softcover, 2014)|
(You can print this review in landscape mode, if you
want a hardcopy)
Reviewer: Mark Lamendola, author of over 6,000 articles.
This book is very smartly written!
Having been involved in both electrical power monitoring (very data
intensive) and business intelligence software (provides business reports
from database sources) for well over a decade now, I agree with the author's
premise that there's a difference between data and information. I wrote an
article on this subject for the Crystal Reports market, and it's featured on
the Crystalkeen Website. Too much
of what pretends to be "analysis" or "information" or "business reports" is
simply reformatted data and not very useful.
Another premise of this author is that the data analysis function serves
the business, not the other way around. This point is often lost upon those
who are supposed to provide the analysis. Rather than answer business
questions, they just provide analysis. Their thinking, such that it is,
revolves around the idea that they best do their jobs when they can do the
neatest tricks with the analysis system.
These are just two examples of several "wrong thinking" ideas that
Foreman addresses in this book. Because these "wrong thinking" ideas are
pervasive and cause the misallocation of millions of dollars of resources in
the typical large company, this book is worth several thousand times its
cover price for the typical large company. Scale down the cost as you scale
down the enterprise, and the multiplier is obviously less dramatic but still
quite potent. Assuming, of course, the reader grasps what Foreman is saying
and acts upon those new insights.
This assumption has some teeth to it, because Foreman is a very clever
writer. In addition to using humor to keep the reader engaged, he apparently
labored long and hard over his word choices to get clear meaning across to
the reader. This is something I greatly appreciate in a work of nonfiction.
Typically, the subject matter expert lacks such a command of English and
something gets a bit muffed in the translation from text to the mind of the
Now, that's my commentary on the high-level stuff. Which does not
comprise the bulk of this book. I addressed it first because, to me, this
alone makes this book a "must read" for anyone involved in data analysis,
business intelligence, or related fields. Too many in these fields cannot
see the forest for the trees, and their penchant for getting mired down in
insignificant details shows in the results of their work. They wonder why
users waste many hours trying to do their own analysis in Excel, instead of
looking at whether they are providing a useful service to the business and
its decision-making needs.
Let's move on to the technical stuff covered in this book. At one time in
my career, I was a spreadsheet junkie. I built very complex models in Excel.
So I was delighted to walk through Foreman's examples and tutorials on using
Excel to do various kinds of analysis. These examples and tutorials comprise
the bulk of this book, but they are not the point of the book.
Let me explain by analogy. I'm not sure if this reaches the typical
reader, but try to follow (and accept my apologies if it's a dud). In
electrical engineering today, software does the number crunching for you.
But in engineering school (and often in the friendly debates engineers
have), the modus is on manually doing the calculations. When you read the
electrical engineering trade publications, you find not an admonition to run
the example through your software but you find manual calculations being
The reason, in all instances, is the participants must be able to
understand the concepts. You can do this only by crunching the numbers
yourself and following along in the mental processes of arriving at the
answer. So the author of an article might provide quite a trail of
calculation to prove a point. It's the point that matters, not the
calculation per se. But you don't get the point unless you can see how it's
For example, in this book Foreman discusses K-analysis. How can you
really understand this without working through some examples and watching
the effects on the data? Answer: You can't.
To me, being walked through this litany of hard-to-grasp data analysis
concepts is the only way a person can really understand those concepts. I
think a mere surface knowledge is insufficient (a little knowledge is
dangerous....). Even outside the realm of data analysis, people toss about
terms they clearly do not understand but think they do. But based on my many
years interacting with Crystal Reports administrators and trainers, I think
the problem is especially pernicious in this particular field of data
analysis. If you really want to know what you're talking about, you need to
do the learning work.
The first nine chapters walk the reader through data analysis concepts.
Chapter 10 is an introduction to an analysis program called R. Foreman
begins by summing up the previous nine chapters as an exercise in learning
analytics and then making it clear that Excel isn't the right tool for
actually doing analytics.
I don't believe Foreman is trying to "sell" R per se. It's what he's
familiar with. There are other tools for data analysis, including the big
players in the Business Intelligence (BI) market, such as Crystal Reports
and Cognos. Basically, if you want an effective, accurate, efficient way to
answer business decision-making questions from the data your business
gathers, you need to step up to a tool designed for that job. And, of
course, you need an adequate database behind it.
Foreman has excellent advise in his 11th chapter (which is not numbered),
"Conclusion." It's only six pages long, but what he says in here is
profound. If you, as the reader, grasp nothing else but what's in this
conclusion, the book has served you well.