Credit Scoring in Africa: Employing Logistic Regression on Alternative Data

This paper presents an innovative approach to credit scoring for the unbanked population in Africa by leveraging alternative data from Mobile Network Operators (MNOs). Traditional credit scoring methods often fail due to a lack of formal credit history. This research proposes a model that utilises mobile phone usage patterns, such as call frequency, data usage, and mobile money transactions, to assess creditworthiness. A predictive model was developed using logistic regression, a machine-learning technique for binary classification. Initial results show a 90% in prediction accuracy compared to less than 80% when using traditional methods, opening up possibilities for financial inclusion. Key challenges such as data privacy and regulatory compliance are discussed, highlighting future directions in this rapidly evolving field. Despite these challenges, there is an urgency to extend credit services to the unbanked. The potential societal and economic benefits are significant, including new customer bases for financial institutions and improved credit access for unbanked individuals, leading to a more inclusive and prosperous society.