About the project
A Finance demo that follows supplier invoices from received to approved to paid. The data comes first: a seeded Python generator writes a fake accounts-payable export of 1,596 rows from 15 suppliers and six departments, with stories planted for a dashboard to find — Marketing takes about 11 days to approve, logistics suppliers are paid late. It is messy the way real exports are: dd/mm dates, amounts as text, 18 duplicated rows, and two negative amounts that are genuine credit notes, not errors. Next come a Power BI model, a Power Automate flow and a Copilot Studio agent.
PythonStar-schema CSVsPower Query (next)Power BI (next)