Good decisions
start with data
you can trust.

Using fixed synthetic 2025 UK fashion data, I built tested dbt and DuckDB models, governed metrics through Wren AI, and delivered a browser dashboard for seven-day last-non-direct attribution and monthly campaign analytics.

A fictional digital wallet business needs reliable visibility into transaction performance, customer activity and retention. I built a tested dbt and BigQuery reporting layer that turns raw wallet activity into consistent product KPIs, including GTV, success rate, DAU, MAU and retention.

A fictional healthcare organisation needs better visibility into appointments, patient flow, admissions, readmissions and facility capacity. I built a tested dbt and BigQuery reporting layer that separates appointments, encounters and admissions and produces operational KPIs such as no-show rate, wait time, length of stay and bed occupancy.

An e-commerce business needs dependable reporting across orders, products, sellers, payments and customers without distorting results across one-to-many relationships. I built a Snowflake and dbt Cloud pipeline that converts nine source tables into tested dimensional models for reliable order, product and seller reporting.

Order totals can change when payment joins duplicate rows or currency is converted twice. Using the Jaffle Shop sample dataset, I built a Wren semantic layer over a dbt-built DuckDB warehouse to make those rules explicit. I reconciled order and payment amounts and checked that the monthly view and metrics cube agreed. The public dashboard contains saved, independently verified example answers.