
Credit Default Risk Analysis for Nova Bank
I applied Random Forest and XGBoost to credit risk data to uncover how borrower traits, loan details, and credit history influence default risk across the USA, UK, and Canada.

Click to Conversion: Tracking Paid User Behavior
This project analyzes user behavior before purchase by mapping paid users’ session journeys using SQL views and event tracking data.

Truth or Trick: Deep Learning for Fake News Detection
This project builds a fake news detection model using LSTM-based deep learning and GloVe word embeddings to classify news articles as real or fake.