Showcase
Five calculations, run end to end on public data.
Each one goes from raw data to a payout report, and each report says where its method stops. They are proofs of concept on public data, not live policies.
At a glance
What each one shows, and where it stops.
| Calculation | Peril | Data | Backtest against losses |
|---|---|---|---|
| La Réunion | Marine heatwave · coral bleaching | NOAA Coral Reef Watch, 5 km Degree Heating Weeks | scored 83% hit rate |
| Mayotte | Marine heatwave · coral bleaching | NOAA Coral Reef Watch, 5 km Degree Heating Weeks | scored 100% hit rate |
| Auckland | Extreme rainfall · pluvial flood | NASA IMERG satellite rainfall | not scored no loss series |
| New Zealand | Tropical cyclone · wind | ERA5 reanalysis, 10 m wind gust | not scored no loss series |
| Bangladesh | Agricultural drought · rainfed rice | CHIRPS rainfall and MODIS NDVI, blended 50 / 50 | not scored no loss series |
Rainfall indices: CHIRPS v2 production ends after December 2026; new work uses CHIRPS v3. The Bangladesh run uses CHIRPS v2 and would be re-backtested on v3 before any contract.