I read the data
so the decision
doesn't have to guess.

End-to-end capstone joining Python EDA with a Random Forest churn model and a four-page Power BI dashboard across 10,000 bank customers. Age and account balance surfaced as the strongest churn predictors — the model correctly calls 9 in 10 customers.

Four-page interactive dashboard analyzing IBM HR data across 1,470 employees. Sales Representatives logged the highest attrition at 39.8%, with employees under 25 leaving at nearly double the company average.

Four-table star schema queried with JOINs, window functions and aggregate business logic. Found Furniture sales volume masking a net loss caused by 20%+ discounting — the kind of finding that only shows up when you query, not just visualize.

Linear Regression model predicting residential sale prices on the Ames Housing dataset. Overall quality rating and above-ground living area carried the strongest correlation to final sale price.

K-Means clustering identifying 5 distinct mall customer segments by annual income and spending score, with optimal cluster count determined through the Elbow Method rather than guessed.

Exploratory pass on IBM HR data preceding the Power BI build — identifying that overtime status correlates more strongly with attrition than department or job role alone.

End-to-end capstone joining Python EDA with a Random Forest churn model and a four-page Power BI dashboard across 10,000 bank customers. Age and account balance surfaced as the strongest churn predictors — the model correctly calls 9 in 10 customers.

Four-page interactive dashboard analyzing IBM HR data across 1,470 employees. Sales Representatives logged the highest attrition at 39.8%, with employees under 25 leaving at nearly double the company average.

Four-table star schema queried with JOINs, window functions and aggregate business logic. Found Furniture sales volume masking a net loss caused by 20%+ discounting — the kind of finding that only shows up when you query, not just visualize.