Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die Hardcover – January 1, 2013

Dr. Siegel seems to have written this book for those with limited math skills, but with a desire to better understand the techniques for extracting meaning from big data. Since this describes me, I found the book quite valuable and gave it my highest rating. If you already have a strong grasp of the tools for organizing and interpreting big data, the book will probably not meet your needs.While the author writes well, the Introduction and first chapter skipped around on topics and anecdotes which caused me some initial concern. However, keep going because once past this early stage, the book gained traction quickly. In chapter 2, the author considers ethical concerns arising from predictive analysis. Target's analysis of a woman's buying patterns for pregnancy and Hewlett-Packard's analysis of its own employees for those that may quit both raise thought-provoking issues of whether such analyses are, to use the author's phrase - insight or intrusion. Using predictive analytics to prevent online fraud probably isn't as controversial.The author describes the tools to undertake predictive analysis. Decision trees, ensembles, and ensembles of ensembles may all be used to draw meaning from data. He describes the IBM team's development of Watson for the famous contest on the game show "Jeopardy" when Watson beat two humans who had performed at championship levels on this show. The author details the challenges of natural language processing to enable a machine to derive meaning from spoken English. He goes through examples that illustrate the high-level challenge. The IBM team used the tools of ensembles of ensembles (read the book to understand this) coupled with statistical interpretation to determine the most likely correct answer to any question and to do so faster than the human contestants. This was machine learning at the currently highest level. One of the fascinating points is that art drives machine learning.Can predictive analytics be employed to forecast an individual's actions? The thought seems troubling to me, but the possibility that prediction could be so used must be recognized. Which persons are most likely to favorably respond to a cell phone renewal offer as opposed to interpreting the offer as an opportunity to seek another carrier could have meaningful financial implications to a telecommunications carrier. He describes a predictive modeling undertaken by Oregon to predict which potential parolee is more likely to commit another crime if released from prison. This, too, has real world implications for the potential parolee and society. Once again, predictive analytics challenged me on many levels.Dr. Siegel identifies five effects of prediction. These are: (1) the prediction effect; (2) the data effect; (3) the induction effect; (4) the ensemble effect; and (5) the persuasion effect. I encourage you to read the book to learn about these effects and consider their potential cumulative "effect" on society, for good and ill.

Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die is a must read for anyone and everyone, technical and non-technical, that is serious about wanting to improve their business. It explains the amazing potential and power of predictive analytics in a truly enlightening, entertaining, and witty fashion. Fern Halper, the Director of TDWI Research, said in a recent article, “predictive analytics is a technology whose time has finally come”. This book takes a giant leap in making that statement true by taking us from 0 to 60 mph about what PA is and how it works and by providing a plethora of dramatic examples of how this technology has already been successfully used in a wide variety of industries. My favorite witticism in the book is: “The elephant in the room is that there is no elephant in the room.” What a great way to help us put big data into perspective!Jerry SabudaPresident, Sabuda Technology Solutions LLCPittsburgh PA

Clear and concise, this book is a good introduction to predictive analytics... but not only for marketers (although if you're one of them, you probably should read this book). From insurance companies and financial to every sector and company, no one can escape to the temptation to analyze and predict the future, but thankfully in a more "empirical" way: understand the things that happen and the reasons why they happen.This book is not and end in itself. It just open the door to a (ok, maybe now not so) new approach, giving you the tools to see and understand what happens around us under new eyes.

This book is interestingly written to generate excitement in predictive analytics using some recent successful application. If the book is targeted to general public, then I think this is still quite technical, if this is directed towards practitioners or business managers then it falls short in describing suitable applicability parallels. The writing style is quite good though the author is sometimes given in to hyperbole..The reference section is quite exhaustive and valuable. The author may have generated more excitement by adding a more "How to chapter.." Folks in the industry who grapple with data everyday need clean linear road maps with respect to tools and applicability of techniques and the books does nothing of that sort.With recent media reports on data security etc., I expect in a future edition issues around depersonalization, storage etc to be talked about as well.I had asked a friend of mine to carry the book to India on one of his business trips as I was not able to find this here. And waited with bated breadth for it to arrive. But overall it was bit of a damper....

This book is aimed mostly at people who are interested in learning about where (as opposed to how) one can effectively use Predictive Analytics and related technologies such as Machine Learning and Natural Language Processing. There is some high level discussion of algorithms such as linear regression, decision trees, random forests and even a nice discussion about Watson's question answering algorithms. The book has many examples of where Predictive Analytics can and is being used. Some of these are relatively obscure, because companies prefer to make money off these techniques rather than talk about it (and dilute their competitive edge). The narrative is interesting and humorous, and the author shares many anecdotes from his own life, having lived through Predictive Analytics relatively short life-span. Finally, the bibliography/reference section lists URLs that will probably take you months to get through. All in all, a "popular" book aimed at people who are looking into learning about and/or adopting Predictive Analytics rather than established practitioners, but very useful and well written nevertheless.