Machine learning project that predicts individuals’ access to formal financial services in Ghana using the Global Findex dataset (143K+ records, 128 variables). Built logistic regression models with feature engineering, regularization, and K-fold cross validation, achieving 81% accuracy