Agriculture
Know what the season did to your crop before the harvest tells you.
Growers, lenders and insurers make capital decisions on estimates that lag the field by weeks. We compress that lag to days, at scale, and show the error bars.
Use cases
Weather-event impact assessment
Frost, flood and drought damage mapped field by field within days of the event, with severity classes and a reconciliation against reported losses.
Pruning and management-event detection
Detect when blocks were pruned, hedged or harvested from satellite time series, so operations and forecasts stay in sync.
Yield forecasting
Season-long forecasts from vegetation-index integrals, updated as the season develops, calibrated to your ground truth.
Collection programme design
Where to send scouts, drones and vehicles; what to capture; how often. Built to the budget, not to the technology.
Data we typically use
- — Sentinel-2, Landsat 8/9, PlanetScope
- — Sentinel-1 radar
- — Daily weather and land-surface temperature
- — Drone orthomosaics and vehicle-mounted cameras
- — Field capture (ODK / Kobo-style forms)
What you receive
- — Field-level impact maps and severity tables
- — Forecast reports with validation and error bars
- — Collection plans with cost, cadence and sensor per zone
- — Interactive map, GeoPackage / GeoParquet, PDF map series
Work
Questions we get
- Do we need to buy commercial imagery?
- Often not. Public archives cover most field-scale questions; we recommend commercial or drone capture only where the resolution changes the decision.
- How do you validate?
- Every analysis ships with a validation table — against your records where you have them, against public figures where you don't.