Retail Banking Transaction Analysis
Designed a relational banking database with primary and foreign keys; analyzed customer behavior, credit/debit trends, spending categories, and branch-level transaction patterns using SQL, including JOINs, GROUP BY, HAVING, aggregate functions, filtering, sorting, and subqueries.
UPI Fraud Detection using CNN and LSTM
Developed a deep learning model to detect fraudulent UPI transactions using CNN and LSTM; performed data preprocessing, feature scaling, and EDA; evaluated performance using Accuracy, Precision, Recall, F1-Score, and ROC-AUC. Technologies included Python, Pandas, NumPy, Scikit-learn, TensorFlow/Keras, Matplotlib, CNN, and LSTM.