12 Iterating on someone else’s work: Data preparation

 

This chapter covers

  • Continuing the work of another analyst
  • Investigating and verifying an existing analysis
  • Preparing event-level data to be suitable for user-level segmentation

Every analyst will need to continue someone else’s work at some point. This “someone” might be a past version of you from months ago. The process for working on the second version of a project is the same as starting from scratch.

Because we will own this new version, we still need to understand the problem, look at the available data, and so on, even if someone else has already done it. In this project, you will get the opportunity to practice taking over from someone else. Another analyst has prepared the minimum viable answer to a stakeholder question, on which you will iterate.

The specific topic of this chapter is one that is also common in the real world: segmentation. Most businesses have questions in the form of “How are some things similar to other things?” where the thing in question could be anything from a product to a customer to an entire geographic area.

12.1 Finding similar entities

12.2 Continuing someone else’s work

12.3 Project 8: Finding customer segments from mobile activity

12.3.1 Problem statement

12.3.2 Data dictionary

12.3.3 Desired outcomes

12.3.4 Required tools

12.4 Applying the results-driven method to creating the second iteration of a customer segmentation

12.5 An example solution: Creating customer segments

12.5.1 Recreating someone else’s analysis

12.5.2 Analyzing event data to learn about customer behavior

12.5.3 Project progress so far

Summary