BASIC STATISTICS OF CUSTOMER ANALYSIS DATA IN THE TELECOMMUNICATION SECTOR

Abstract

The telecommunications sector has reached a broad customer base with the increased use of smartphones. However, challenges such as customer churn have also emerged in this sector. This study encompasses a statistical analysis to develop strategies to prevent customer churn by analyzing customer behavior in the telecommunications sector. Basic descriptive and inferential statistics were applied to Ghana's telecommunications sector dataset. The dataset includes demographic information, service usage trends, and customer statuses of 1971 customers from four companies.

The study evaluated factors affecting customer churn, such as age, gender, education level, duration with the company, payment amount, and reasons for choosing the company. The customer churn rate was found to be 52.5%, with churns being particularly concentrated in the 25-34 age range. Additionally, pricing was identified as the most significant reason for churn, and it was observed that churn rates increased among long-term customers (3-4 years and above).

As the primary motivation of the study was to prepare a dataset for customer churn prediction using machine learning methods, the preprocessing, cleaning, and basic statistical analysis phases of the data were detailed. Furthermore, the findings provided a foundation for advanced analyses and strategic decisions aimed at enhancing customer satisfaction and reducing churn. In this context, the research results offer significant contributions toward the development of customer-focused strategies in the telecommunications sector and the more effective use of machine learning algorithms.