Overview

In this practical we are going to analyse a small churn dataset and quickly build a (baseline) classification model. In the subsequent weeks, we will revisit various steps along the pipeline and hopefully improve our resulting model.

Churning is the movement of customers from a company in favour of another. This is of significant interest to mobile providers and utility companies. Typical aims of a churn dataset is to:

Develop models that predict whether a customer is likely to churn, i.e., a company would like to know how to characterize the customers who may soon churn.
Moreover, based on the knowledge gained from analysing and mining the data, suggest practical measures that the company can put in place to try to retain those customers likely to churn.

When analysing a dataset, you should follow the KDD process: data understanding (usually called exploratory data analysis), data preprocessing, mining the data, and description of the interesting discovered patterns (knowledge).

Published work that used this Dataset

You might find it of interest to see how others have studied a similar churn dataset.

A Neural Network based Approach for Predicting Customer Churn in Cellular Network Services