Data‐Intensive Analysis
Óscar Corcho, Jano I. Van Hemert · 2013
This chapter presents a set of scenarios that introduce different types of data analysis problems of increasing complexity, which could be handled by a team of data analysts in a telecommunications company, Telco Inc. It starts with a rather conventional scenario where a customer churn prediction model is built from the company's own data sources. Then, it moves to a more complex scenario where, following a company merger, the data sources of two companies have to be integrated, increasing levels of heterogeneity in the generation of the churn prediction model. Finally, it describes a scenario where less conventional techniques—graph analysis and stream mining—have to be applied in a more-heterogeneous and less-controlled scenario, with data originating not only from company databases but also from streams of social network information, such as microblog feeds. Controlled Vocabulary Terms multivariate statistics