CHAPTER 4 Descriptive Analytics

In descriptive analytics, the aim is to describe patterns of customer behavior. Contrary to predictive analytics, there is no real target variable (e.g., churn or fraud indicator) available. Hence, descriptive analytics is often referred to as unsupervised learning because there is no target variable to steer the learning process. The three most common types of descriptive analytics are summarized in Table 4.1.

Table 4.1 Examples of Descriptive Analytics

Type of Descriptive Analytics Explanation Example
Association rules Detect frequently occurring patterns between items Detecting what products are frequently purchased together in a supermarket context Detecting what words frequently co-occur in a text document Detecting what elective courses are frequently chosen together in a university setting
Sequence rules Detect sequences of events Detecting sequences of purchase behavior in a supermarket context Detecting sequences of web page visits in a web mining context Detecting sequences of words in a text document
Clustering Detect homogeneous segments of observations Differentiate between brands in a marketing portfolio Segment customer population for targeted marketing


In this section, we will address how to mine association rules from data. First, the basic setting will be discussed. This will be followed by a discussion of support and confidence, which are two key measures for association rule mining. Next, ...

Get Analytics in a Big Data World: The Essential Guide to Data Science and its Applications now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.