Hands-On Prescriptive Analytics

Chapter 1. An Analytical Framework

All books—any written document, in fact—must start somewhere. The authors of many technical books I have read immediately present the technicalities of their subject. They provide neither background on their subject’s evolution nor motivation for reading their book. I will not start this way. Instead, I will give you my perspective on why Prescriptive Analytics exists and developed, as well as why it is important for data science and business itself. It exists to help decision-makers make decisions.

I will also present a framework for Prescriptive Analytics, since many analytical methodologies are inapplicable for all decision problems. There is no one-size-fits-all. There are different levels of decision problems with some methodologies applicable to one level, but not another. I will later refer to these levels as scale-views.

My main theme is that decision-makers at all scale-views must make decisions that have effects in the future, whether the future is literally tomorrow, next week, next year, or the next decade. They are made under uncertainty simply because no one knows what will happen tomorrow.

The decision-makers ask their data science team (DST) for information to help them make these decisions. The information comes from Descriptive Analytics and Predictive Analytics in the form of a decision menu of options. Unfortunately, there is almost no guidance as to which menu option is best. This adds another layer of uncertainty: What is ...