Semester at Sea — financial data pipeline

An end-to-end pipeline over multi-year institutional revenue data: extraction, schema normalisation, validation, and a SQL modelling layer that turned an inconsistent historical record into something a finance team could actually make decisions from.

What it produced

• Cross-fiscal revenue quantified and reconciled across years that didn't previously line up
• Bad-debt exposure modelled across student cohorts, isolating where the risk concentrated
• A repeatable pipeline, not a one-off spreadsheet — the analysis reruns when the data updates

Unglamorous plumbing, decision-grade output. Most of the value in analytics work sits in the half nobody photographs.