Published sample conversions
Each sample is a synthetic workbook built to contain the habits that make real models hard to migrate: plugs typed over formulas, lost absolute references, hidden sheets, lookups that miss, errors flowing through totals. The results below are the tool's own evidence packs, published as they came out. Nothing here is a client file.
Asset-finance pricing calculator
A deal-pricing workbook of the kind a lending or leasing team keeps: deal inputs, a rate card looked up two ways, a 72-row amortisation schedule, deal economics with NPV, IRR and a break-even rate, a sensitivity grid and a quote summary.
6 sheets, 641 formula cells in 52 logic blocks.
| Python | 641 of 641 reconciled |
| pandas | 641 of 641 reconciled |
| SQL (DuckDB) | 638 of 641 reconciled |
FP&A month-end model
A small company's monthly P&L model, written the way such models grow in real life: an order log with lookups, a revenue roll-up with whole-column SUMIFS, headcount-driven costs, a P&L with running totals and a dashboard.
8 sheets, 2,477 formula cells in 43 logic blocks.
| Python | 2,477 of 2,477 reconciled |
| pandas | 2,477 of 2,477 reconciled |
| SQL (DuckDB) | 2,470 of 2,477 reconciled |