2 Encoding geographies

 

This chapter covers

  • How to use the results-driven approach to tackle a real problem
  • Making analytical decisions in the face of uncertainty

Now that you have been introduced to the results-driven approach, let’s apply it to a real data science task. We will tackle a real-world problem and apply the results-driven method to it. It is a template for approaching the other projects and any other project you encounter in the future.

All the projects and data are available at https://davidasboth.com/book-code for you to attempt them yourself. There, you will find both the dataset for the project and the same example solution presented in the chapter in the form of a Jupyter notebook.

A note on Jupyter notebooks

Jupyter is a development environment that allows the mixing of code and accompanying text in the form of notebooks. Notebook files let you read the text and run the code directly within the document. They are popular for education and data analysis.

The notebooks accompanying the book collect the code snippets from each chapter into a single document where the code is also runnable. Sites such as GitHub even let you view the notebooks in the browser without requiring Jupyter to be installed. See the appendix for instructions on how to install Python to run the code yourself.

For this chapter’s project, let’s examine an example task that analysts encounter, which at first glance might seem simple enough—encoding geography.

2.1 Project 1: Identifying customer geographies

2.1.1 Data dictionary

2.2 An example solution: Finding London

2.2.1 Setting ourselves up for success

2.2.2 Creating the first iteration of a solution

2.2.3 Review and future steps

2.3 How to use the rest of the book

Summary