Predictive Analytics for the Modern Enterprise

Chapter 7. Predictive Analytics for Business Problem-Solving

So far we have talked about the foundations of predictive analytics. We have talked about algorithms and how they work, libraries and tools that are available for working with predictive analytics, and how to set up these libraries and tools and use them with generic data. In this chapter, we will discuss three use cases, each from a different vertical industry, to see how predictive analytics can be applied in the field. The purpose is to look at scenarios that represent business problems and how predictive analytics can be applied to solve them. As we go through this process, we will cover data preprocessing, specifics of various configurations, and the reasons behind some of the decisions we make as part of building enterprise models for predictive analysis.

Prediction-Based Optimal Retail Price Recommendations

In the retail industry, one of the key metrics for success is revenue. Several techniques are employed across large retail chains to maximize revenue from sales, and the enterprise usually has control over the selling price of a commodity within a particular price range, which can vary by product and region. For example, some cities follow the Manufacturer’s Suggested Retail Price (MSRP), while others allow selling above or below that price. In the United Arab Emirates, for instance, gasoline prices are fixed countrywide each month; in the United States, prices can vary from one state to another. Additionally, ...