Remote sensing methodology, Sentinel-2 processing techniques, crop insurance actuarial applications, and commodity trading signal construction — written by the Terraxiq data science and markets team.
County-average loss-cost factors were a reasonable approximation in 1980. Forty years of climate variability and land-use change have made them a liability for underwriters writing sub-county policies.
Normalized Difference Vegetation Index wasn't designed for insurance underwriting. Here's how it became one of the most reliable early-season yield proxies available.
10-meter vs. 30-meter resolution sounds like a technical footnote. When you're delineating individual field polygons for insurance pricing, it's the difference between accurate and wrong.
USDA publishes monthly. Commodity markets move daily. The 30 days between WASDE publications create a structural information asymmetry — and satellite yield estimates are closing it.
The Midwest gets cloudy in July exactly when crop stress signals matter most. Temporal compositing across multi-day satellite passes solves the coverage problem — here's the methodology.
The shift from county-unit to enterprise-unit policies in crop insurance requires field-boundary precision that government survey data can't provide. Here's what satellite-derived boundaries change.
Machine learning yield models are only as good as their calibration data. Our process uses USDA NASS county-level yield surveys as ground-truth holdout — here's what that looks like.
Actuaries are not GIS analysts. A geospatial data product that requires ESRI expertise to operate will never reach the underwriting desk. Here's how we designed our API for non-GIS consumers.
A single-number yield forecast creates false precision. Probability distributions let traders quantify the downside and model position sizing across yield scenarios — this is how they're built.
NDVI captures what plants look like from orbit. Growing degree days capture when development stages occur. Combining both signals produces materially better mid-season yield forecasts.
Winter wheat matures under cloud cover and snow. Optical satellite imagery fails exactly when Kansas and Colorado hard-red wheat conditions matter most. How radar fills the gap.
Loss ratios in US crop insurance have averaged above 0.90 for parts of the past decade. Sub-county actuarial precision enabled by satellite field data offers a structural improvement — here's the math.
