Hafsa Dania User

I investigated 96,000+ delivered orders across five integrated datasets to identify a 32.3% revenue decline over three consecutive weeks. Using SQL, Python, statistical testing, and Power BI, I isolated seasonal category demand slowdown and quantified ₹2,46,430 in revenue loss.

I built a reproducible SQLite analysis using multi-table JOINs and conditional aggregation, then trained a Random Forest model to identify freight value and product category as the strongest revenue drivers. I also delivered an interactive three-page dashboard so non-technical stakeholders could explore weekly trends, category performance, and seller insights independently.

For an e-commerce churn and retention project, I validated 17 predictors, engineered five customer-behaviour features, and used cohort analysis to identify the first six months as the highest-leverage retention window. My Logistic Regression and Random Forest models achieved 0.9985 ROC-AUC and 96% churn recall, informing a three-tier retention strategy projected to recover ₹2,98,800 in revenue.

I'm a Computer Science and Engineering student with hands-on experience in SQL, Python, Power BI, data cleaning, statistical analysis, machine learning, and stakeholder-ready dashboards.