CONCEPTION ET IMPLEMENTATION D'UNE INFRASTRUCTURE CLOUD PRIVE INTEGRANT UN DATA WAREHOUSE ET DES TECHNIQUES DE MACHINE LEARNING POUR L'ANALYSE DES DONNEES STRATEGIQUES DES ABONNÉS: CAS DE VODOCOM CONGO (RDC).

The continuous growth of data generated by telecommunications operators has become a major challenge for strategic analysis and decision-making. However, fragmented information systems and the lack of integrated analytical infrastructures often limit organizations’ ability to transform data into actionable knowledge. In this context, this study proposes a private cloud computing infrastructure integrating a Data Warehouse, Machine Learning techniques, and Business Intelligence tools for the analysis of strategic subscriber data at Vodacom Congo (DRC). The proposed architecture is based on VMware ESXi virtualization, a SQL Server Data Warehouse, the K-Means algorithm for subscriber segmentation, and Discriminant Analysis for churn prediction. Experiments conducted on a dataset of 764 subscribers identified three distinct customer segments and achieved an overall classification accuracy of 90.8%. The results demonstrate that an integrated approach combining Private Cloud Computing, Data Warehouse, Machine Learning, and Power BI significantly enhances customer insight, behavioral prediction, and decision-making processes within telecommunications companies.