Preface

Predictive analytics is one of the most important technologies of our time. Every day, companies in all industries and all types of institutions are using predictive techniques to solve a wide range of problems. Although many of the main ideas and techniques have been around for many decades, the use of predictive analytics has exploded recently due to the increased ability to capture and store data, which is the raw material from which we build predictive models. There are two other big factors that explain the increasing adoption of this technology: the first is the astonishing increase in computing power, and the second is the availability of many open source software projects that have given access to professionals outside academia to many of the most powerful predictive analytics techniques. The Python programming language and its ecosystem of analytical libraries, also known as Python's data science stack, is such a project and has democratized the use of advanced analytical techniques. 

This is a book about predictive analytics, but rather than focusing exclusively on explaining in detail the algorithms and techniques, this book is more about the process of doing predictive analytics in the real world. The main goal of this book is to make you familiar with all the stages in the process of solving a business problem using predictive modeling and to show, with hands-on examples, how to use Python and its data analytics ecosystem to implement many of the main techniques and approaches used in real-world predictive analytics projects. We use two main projects in this book and walk you through the entire predictive analytics process: from business and problem understanding to model deployment, all through hands-on examples.

There are many techniques that can be used for predictive analytics: statistical models, time series analysis, and spatial statistics to mention a few; however, in this book, we focus on the most widely applicable and successful set of techniques: machine learning, specifically the branch of supervised learning.

In my view, a predictive model is only a means to an end. The goal of using predictive analytics is to solve problems; therefore, a good predictive model is not one that uses the latest and most fashionable techniques, nor is it the most complicated or the simplest one. A good predictive model one can be used to solve a real-world problem in a satisfactory way. My goal is that by the end of this book you will have the foundations that you need to start solving real-world problems using predictive analytics.