Credit Card Fraud Detection
Finished my Credit Card Fraud Detection project by simulating real-world protection for banking transactions in Poland. What I did: • Processed the ULB fraud dataset (285k rows, heavy imbalance) • ETL → PostgreSQL (Docker), feature scaling + PLN adjustment • EDA → identified key fraud signals (V14, V17 strongest) • Balanced data with SMOTE → trained RandomForest & achieved a 0.99 AUC, high recall • Built interactive Tableau dashboard (KPIs, anomalies, correlations) • Deployed live prediction app in Streamlit Learnings: full pipeline ownership (ETL → modeling → BI → deployment), imbalance handling, explainable ML, stakeholder visualization. Find the Live demo on Streamlit here: https://creditcardfraud-7mtwnvnteisd2cgcr8p2jf.streamlit.app Open to feedback from fraud/risk, fintech and data science communities! #FraudDetection #MachineLearning #DataScience #Streamlit #Tableau #Python #FinTech
