13 Iterating on someone else’s work: Customer segmentation

 

This chapter covers

  • Segmenting users based on behavioral data
  • Evaluating the output of a clustering algorithm

In this chapter, we will use the dataset created in the previous chapter to segment users into groups based on their mobile browsing behavior. In chapter 12, we took over another analyst’s work, verified their findings, and continued analyzing our customers’ behavior. We turned an event-level dataset into a user-level one, and in this chapter, we will apply a clustering algorithm to create distinct user segments. We will evaluate these segments and address our stakeholders’ questions. Let’s recap the project before continuing.

13.1 Project 8 revisited: Finding customer segments from mobile activity

To recap, we’re working for AppEcho Insights, an analytics company focused on mobile user behavior. They analyze data on how users use their phones and provide insights to mobile phone manufacturers and app developers.

They want to understand whether there are groups of users who use their phones in a similar way. Knowing these user segments would be useful as their clients could target entire user bases with different initiatives. They’d like to focus on the following:

13.1.1 Data dictionary

13.1.2 Desired outcomes

13.1.3 Project summary so far

13.1.4 Segmentation of mobile users using clustering

13.1.5 Conclusions and next steps

13.2 Closing thoughts: Segmentation and clustering

13.2.1 Skills learned to use for any project

Summary