Banking Dashboard – End-to-End Data Analysis using Power BI
Built an interactive multi-page Banking Dashboard with Home, Loan, Deposit, and Summary pages. Performed data cleaning, transformation, exploratory data analysis, and managed a 24-column dataset in MySQL. Standardized categorical variables and created Income Band categories in Power BI.
Transaction Fraud Detection - End-to-End ML Pipeline
Developed a fraud detection machine learning model, performed feature engineering and exploratory data analysis on highly imbalanced data, used Boruta for feature selection, and compared seven models. Selected XGBoost with 94.4% precision and 82.9% recall and deployed it as a Flask API.
Smart Disaster Management - Real-time Response Dashboard
Developed a real-time dashboard for active incidents including fires and floods, with auto-dispatch of emergency resources. Used Dijkstra’s algorithm for shortest-route computation, built a disaster and resource-state simulation engine, and added dynamic incident priority escalation.