Big Data @ Work: Dispelling The Myths, Uncovering The Opportunities, by Thomas H. Davenport

The author, as one of the people responsible for popularizing analytics through books and consulting, has a fair amount of goodwill in writing about the phenomenon of big data. Here he makes some sound points about the sort of people who make good data scientists [1], and in looking at the qualities, I saw that I have skills in about three of them, including a generally experimental approach to life as well as strong communication skills. As for the rest, he talks a lot about the technology that makes data comprehensible and is able to talk about both big data and small data, big companies and small companies, with a great deal of authority. In reading the book, I was struck by the fact that I had been one of the ground-level people involved in one of the big data projects that the author talks about, the UPS Orion project, and it was nice to be able to relate to some of the book’s case studies from the perspective of someone involved in it as a technician.

In terms of its contents, the book is about 200 pages long, divided into eight chapters (and a survey which serves as the appendix), most of which are attempting to convince business leaders to take quantitative analysis seriously. Some of the data, surprisingly, is a bit thin [2], missing percentages or other units of measurements on some of the graphs. Also, readers of this book would be advised that reading this book while playing the “Hadoop” drinking game would be extremely unwise, as it is used seemingly dozens of times per chapter to refer to massive and inexpensive big data storage solutions. A hadoop, after all, is a data server solution that involves splitting up overwhelmingly large data into many servers so that it can be analyzed faster. The author refers to this solution many times in this book, sometimes ironically and most of the times glowingly, making it an obvious solution for companies looking to get a handle on large amounts of data. Certain companies, LinkedIn in particular, come off looking very skilled at dealing with data, and the author’s conclusions about more meritocratic organizational structure will find no argument here. This book has a lot to like.

In many ways, this book sits at the intersection between two trends at cross purposes with each other. On the one hand, this is a book that celebrates quantitative analysis, often done by people with an odd match of skills between effective written and spoken communication and a passionate interest in solving problems via data and experimentation rather than relying on human intuition. Yet on the other hand, this is a book written in the language of a business consultant who is writing about something that is overhyped but still important, seeking to convince business leaders who themselves are not particularly interested in big data or quantitative analysis in any form, except that it helps to better the bottom line of their companies. The fact that this is not done directly, but indirectly through an appeal to the positive examples of companies ranging from UnitedHealth to the University of Pittsburgh makes this a more subtle achievement. As far as a read goes, it was more than a little salesy, but all the same, it was about a worthwhile subject, written by someone who clearly knows what he is talking about and writes with wit and humor as well as considerable background knowledge.

[1] See also:

https://edgeinducedcohesion.wordpress.com/2014/08/26/fragmented-book-reviews-part-four/

https://edgeinducedcohesion.wordpress.com/2015/06/30/book-review-design-for-information/

https://edgeinducedcohesion.wordpress.com/2015/04/09/im-just-here-because-i-love-data/

[2] Not quite up to the level of Tufte, but at least superior to this:

https://edgeinducedcohesion.wordpress.com/2015/07/30/book-review-rise/