Our first guest for the series is Dr. Eric Siegel. Eric is the founder of Predictive Analytics World (PAW<\/span><\/a>). PAW is a business-focused event for predictive analytics professionals and practitioners.\u00a0 Moreover, Eric is also the author of the book ‘Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die’<\/span><\/span><\/span>. Amazon has ranked Eric’s book as the #1 Amazon category bestseller. I personally divide books on analytics into a couple of categories i.e.<\/p>\n Eric’s book perfectly fits the bill for the second category. It is a really well written book with tons of real world examples of predictive analytics. I had a great time while reading this book. Eric has done a phenomenal job of making predictive analytics accessible to the general audience. Additionally, experts of the field will have several wow moments while reading this book.<\/p>\n I had an opportunity to ask Eric some questions regarding predictive analytics and the future trends in the area. Here are exerpts from my conversation with him.<\/p>\n Roopam Upadhyay:<\/strong> <\/span>Hi Eric, congratulations on your successful book \u2018Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die<\/em>\u2019 and thanks for taking the time to talk to YOU CANalytics. Did you predict the kind of success your book has achieved?<\/p>\n Eric Siegel:<\/span> <\/strong>Thanks! Well, while writing, I certainly knew there should<\/em> be demand for an accessible introduction to predictive analytics<\/strong> for all readers — it’s an important topic! I’ll also mention that — for those already well versed in the topic — the book also covers case studies and (in later chapters) advanced topics.<\/p>\n As far as predictive analytics literally predicting itself, there are actually some examples, such as predictive analytics software vendors using their own tool to target the marketing of itself.<\/p><\/blockquote>\n Roopam:<\/strong><\/span> You mentioned in your book that predictive analytics is really difficult to explain to newcomers. However, you seem to have mastered the art of explaining complex subjects in easy-to-read ways. In fact, even when you were a professor of computer science at Columbia University, you used to compose songs and sing to the students to make the subject fun! Over the years, what strategies have you evolved to explain analytics to a general audience?<\/p>\n Eric Siegel:<\/strong>\u00a0<\/strong><\/span>In my book’s preface, I do make light of the challenge of describing what I do to new friends, relative to the much easier time lawyers, waiters, and doctors have answering the question, “What do you do?”<\/p>\n Predictive analytics is easy to inadvertently explain poorly <\/strong>to newcomers. It is important to lead with its value – which comes of driving many operational decisions via many individual per-individual predictions – and also describe how these predictions are derived by way of learning from data… which, in turn, is actually an encoding of experience. So there are many levels that must be hit before a newcomer has the complete, basic picture.<\/p><\/blockquote>\n Roopam:<\/strong> <\/span>Let\u2019s talk about careers. You did your Masters and doctoral degrees in computer science at Columbia University and joined the same university as a faculty, you got teaching awards. What motivated you to switch from the academic life to starting out on your own?<\/p>\n Eric Siegel:<\/strong>\u00a0<\/strong><\/span>I find the challenges of deploying this technology originally borne of research labs, predictive modeling,<\/em> just as interesting as the inner workings of the analytical methods themselves. It’s exciting to see these great ideas come to fruition as real value when deployed!<\/p><\/blockquote>\n Roopam:<\/strong><\/span> I could completely understand the joy of deploying predictive analytics solutions and reaping fruits of labor. What tips would you like to share with someone starting out in predictive analytics and data mining?<\/p>\n Eric Siegel:<\/strong><\/strong>\u00a0<\/strong>Within each project, focus first on the business objective. Specifically, what operational decisions will be driven with predictions, and exactly what is the predictive goal, the target of predictions (i.e., what precise behavior or outcome are you aiming to predict, for each individual)?<\/p><\/blockquote>\n Roopam:<\/strong> <\/span>That’s great advice Eric, business questions or objectives should be the driving force to push predictive analytics projects. What challenges do you see for young entrepreneurs to start their analytics company?<\/p>\n Eric Siegel:<\/strong><\/strong>\u00a0<\/strong><\/span>The biggest initial hurdle is getting your first hands-on project with predictive analytics. It is a bootstrapping issue, because you need it in order to get it. Hopefully, you can find an opportunity to conduct such a project alongside an expert with specific experience.<\/p><\/blockquote>\n Roopam:<\/strong><\/span> \u2018Big data\u2019 seems to be the new buzzword however the term is also quite confusing. What is your definition of \u2018big data\u2019 and where do you see the field going?<\/p>\n Eric Siegel:<\/strong><\/strong>\u00a0<\/strong><\/span>“Big data” is a buzzword effective in generating interest (and hype) that does not introduce any new meaning or substance (it sometimes is used to refer to the specific case when there is so much data, standard DB solutions can’t handle it, but that is a very particular and limited use, and refers in a sense more to an opportunity than a problem). The “big” is really about excitement<\/em>, rather than the absolute quantity of data — the amount of data is growing so quickly, which is really much more pertinent than its current absolute size. Also, “big data” is a grammatically incorrect way to say “a lot of data” — like saying “big water”; it should be, “a lot of water.”<\/p>\n Data is valuable because it is predictive; prediction is the most actionable thing you can get from data. So if there is more data, that’s definitely a good thing, even if you refer to that fact with poor grammar.<\/p><\/blockquote>\n Roopam:<\/strong> <\/span>What are the emerging areas in predictive analytics and data mining?<\/p>\n Eric Siegel:<\/strong><\/strong>\u00a0<\/strong><\/span>I would lead with uplift modeling<\/em> and ensemble modeling<\/em> (covered in 2 of the final chapters in my book).<\/p><\/blockquote>\n Roopam: <\/strong><\/span>Some industries such as banking and retail seem to have a head start in predictive analytics; do you see other industries catching up and how?<\/p>\n Eric Siegel:<\/strong><\/strong>\u00a0<\/strong><\/span>Yes, healthcare and manufacturing are embracing predictive analytics, and we have this year two corresponding inaugural Predictive Analytics World conferences focused on these areas (see the link<\/span><\/a><\/span><\/span><\/span><\/span> for details). It is exciting just how far reaching this technology is, and how quickly it is being adopted across industries. Reflecting this commitment, I have recently been paid to keynote on predictive analytics at conferences within the following industries: marketing, market research, insurance, healthcare, pharmaceuticals, government, business (in general), travel, real estate, and others.<\/p><\/blockquote>\n![\"Eric](\"https:\/\/i0.wp.com\/ucanalytics.com\/blogs\/wp-content\/uploads\/2014\/05\/EricSiegel_hr_headshot.jpg?resize=119%2C179\")
![\"predictive](\"https:\/\/i0.wp.com\/ucanalytics.com\/blogs\/wp-content\/uploads\/2014\/05\/predictive-analytics.jpg?resize=120%2C182\")
Today we are starting a new series on YOU CANalytics called ‘in conversation with’. In this series we will talk to the leaders and experts of predictive analytics and big data to gain deeper insight into the field. <\/span><\/p>\nDr. Eric Siegel<\/span><\/h2>\n
\n<\/span><\/p>\n\n
\n