With RetailPulse AI, I built three Power BI report pages for demand, store and product diagnostics, and forecast comparison, using reproducible Python and SQL metrics.
For the project, I analyzed historical M5 data and made the date scope, missing prices, and data provenance explicit. I also built a local FastAPI and Streamlit demo with a read-only metrics assistant.
I validated an AWS S3, Glue Spark, Parquet, and Athena pipeline using synthetic data, and documented incidents and teardown. The project also included six successful SQL reconciliations.
In my independent data analyst projects, I collected, cleaned, and structured data from more than 10 sources, and applied regression, hypothesis testing, and supervised-learning methods to healthcare-related analysis. Earlier, as a Process & Industrial Safety Consultant, I analyzed manufacturing material consumption; the work reported an 8% reduction in waste and approximately US$450K in annual material savings.
