Part 1: Data Preparation: Business Point of View

Introduction

In Part 1 of this book we will deal with the business point of view of data preparation. This part is free from code and data structures. It contains only the nontechnical components of data preparation. Without reducing the importance of data structures, table merges, and the creation of derived variables, it is fundamental to understand that behind data preparation there always stands a business question.

This part of the book also helps to categorize business questions and to show their impact on data sources and data models.

In Chapter 1 – Analytic Business Questions, we will look at various business questions and will see what they all have in common—that data are needed for their analysis. We will also go through the analysis process and will learn that data preparation is influenced by the business question and the corresponding analysis method.

We will see that a business question itself poses many other questions that need to be discussed in order to prepare and perform the analysis.

In Chapter 2 – Characteristics of Analytic Business Questions, we will show the properties of a business question. We will see, for example, that data preparation is affected by the analysis paradigm, whether it is classical statistics or data mining. We will look at the data requirements, and we will examine whether the analysis is a repeated analysis (e.g., model recalibration over time), or a scoring analysis.

In Chapter 3 – Characteristics of Data Sources, we will go into detail about the properties of data sources from a business standpoint. We will focus on properties such as the periodic availability of data and the advantages of a data warehouse as an input data source.

In Chapter 4 – Different Points of View on Analytic Data Preparation, we will define three roles that are frequently encountered in the analysis process. We will look at the business, analytical, and technical points of view. The different backgrounds and objectives of these roles can lead to conflicts in a project, but they can also provide many benefits for a project if different points of view are exchanged and knowledge is shared.

The intention of Part 1 is to illustrate the business rationale in the data preparation process for analytic data preparation. Part 1 introduces data preparation in general (i.e., from a data modeling and coding point of view). We should keep in mind that all data models, features, derived variables, and tips and tricks we will encounter in this book are provided in order to solve a business problem and are to be seen in the context of the points we discuss in the first four chapters.