This report is composed of three sections. First section discusses several interesting findings in analyzed dataset. Second section focuses on preference-based segmentation. The last section presents classification and discusses effectiveness of classification performed on a general dataset as well as per segment.

As a follow-up to this exercise, we could perhaps conduct a research to assess predictive power of our classification in real setup. In the meantime, please refer to respective sections for more details.

It is recommended to read all sections in the below order.