On my Olist delivery and customer reviews project, I loaded raw data into BigQuery, built documented SQL views, and reconciled the outputs to fixed anchors. My analysis found that 70% of reviews on late orders were written before delivery, changing the interpretation of the review-score gap.
I also built four Tableau dashboards and tested five hypotheses in Python, using statistical methods including ordered logit, piecewise OLS, and inverse-probability weights. In a separate India e-commerce project, I used data-quality checks and SQL views to correct mislabeled orders and identify margin leaks.
At Variable Energy Cyclotron Centre, I optimized experimental parameters against quantitative targets and validated outputs against Monte Carlo benchmarks. As a Visiting Student Researcher at IISER, Kolkata, I analyzed X-ray data using numerical simulation and time-domain analysis.
