quincywambui/Zindi-Challenge---African-Loan-Prediction

# Loan Default Prediction for Financial Markets in Africa ## Introduction Accurately predicting loan defaults is crucial for financial institutions seeking to mitigate risks and optimise their lending decisions. This challenge focuses on building a robust and generalisable machine learning model to predict the likelihood of loan defaults in Africa's rapidly growing financial markets, which are characterised by diverse customer demographics and dynamic economic conditions. The objective of this project is to develop a machine learning model that accurately predicts the probability of loan default, both for existing customers and new applicants. With a particular focus on the African markets, the model will need to generalise well across different countries and economic contexts, as the data for this challenge is sourced from two distinct regions—Kenya (for training and testing) and Ghana (for testing). In addition to prediction accuracy, this challenge encourages innovation by incorporating factors unique to the African financial markets, such as inflation rates, exchange rates, and economic indicators like energy consumption and unemployment rates. By doing so, we aim to create a scalable and effective credit scoring function that helps financial institutions make more informed lending decisions, reduce financial losses, and expand into new markets. ### Key Features of the Challenge: - **Data Sources**: Loan and customer data from Kenya (training and testing) and Ghana (testing), alongside enriched economic data pulled from the Federal Reserve Economic Data (FRED) portal. - **Credit Scoring Function**: Beyond building a predictive model, the top 10 winners will be tasked with designing a credit scoring function that categorises model outputs into risk categories, enabling scalable credit risk evaluation. - **Diverse Market Contexts**: The challenge focuses on developing a model that performs well across different financial and economic environments in Africa. …