Intelligent Customer Segmentation

Overview

In this project the simple rules-based system for segmenting customers was replaced with a clustering method to create more nuanced customer segments. The interface served to help the business user create clusters using custom criteria and explore the contents of the resulting segments.

Summary

Although clustering is capable of creating customer segments which are more nuanced and responsive to the available data, the resulting segments are not defined in human terms. The purpose of this application was to allow the business user to see how clusters compared to each other in order to gain a more intuitive sense of their character. In addition, the segmentation had very different purposes in different parts of the business, so the ability to customize the number of clusters, drop features, and impose rules on the clusters was added in order to allow the business user to experiment with the results and create segmentation that was suited for his or her particular purpose.