Do you use public or proprietary sources?

Yes—both, across several categories.

We ingest a broad mix of global reanalysis archives and national forecast models (e.g., ERA5, MERRA-2, NOAA GFS/HRRR, ICON, GEM, etc.), plus extensive weather-station networks (GHCN-D, ISD/MADIS, METAR/mesonets) and specialty environmental feeds such as maritime buoy and snow telemetry datasets. These provide the physically grounded inputs behind our climate-risk signals.

We incorporate statistics from international agriculture bodies to understand production shares and market structure (e.g., FAO market-share distributions), which lets us interpret climate anomalies in the context of who grows what, and where.

We pull from government price programs/official bulletins and commercial agricultural market feeds (cash and benchmark pricing), giving both standardized reference series and on-the-ground cash quotes.

Alongside these public and licensed inputs, Helios transforms the incoming observations into our proprietary layer of climate risk data, built with proprietary clustering and anomaly-detection methods. Signals are computed on a global mesh of ~4 million hexagons organized into 81 climate profiles, then rolled up to crop/region and market-share–weighted views such as WA%R. The result is a unique risk dataset distinct from the original sources.