Work
Different problems, different builds. Each case study follows the same structure: the problem, the decisions I made, and what I'd do differently.
Engineering Simulation Data Pipeline
A normalized MySQL database and Python ETL pipeline that consolidated simulation data from four scattered sources into one queryable source — with fuzzy matching and change detection.
Olist E-commerce Analytics (dbt)
A 9-table MySQL model of 90K+ orders, rebuilt in dbt Core + DuckDB as a proper star schema — including fixing a grain error that had inflated customer-retention figures.
What these show
Data modeling
Normalized schemas from scratch, star schemas, surrogate keys, and fixing a grain error that had inflated retention figures.
Pipeline engineering
ETL and ELT, parent-child load order, orchestration, automation, reruns that don't duplicate data.
Data quality
Fuzzy matching with confidence thresholds, duplicate and change detection, reconciliation.
Reporting
Power BI, DAX, and data models built for realistic questions.