SeeTree

SeeTree — Block-level perennial-crop yield forecasting corpus with a proprietary satellite foundation model underneath: a cross-industry ground-truth dataset of 600M+ analysed images spanning sugarcane, citrus, oil palm, avocado, almonds, eucalyptus/forestry, coffee, hazelnuts and olives across 12+ countries and every major climate, fused with Sentinel-2 imagery, drone data, weather and grower-supplied agronomic metadata (field boundaries, historical yields, planting and harvest dates). Output is pre-season baselines plus monthly in-season forecasts at block, farm and company level, with API …

SeeTree offers (Geospatial, Alternative, Supply Chain, Fundamental) — 600M+ analysed images across 12+ countries and nine named crops - a stated figure and one of the largest labelled agricultural image corpora held commercially; plus multi-season block-level yield history where growers contributed unlimited historical data, and monthly forecast series per block through each season.

Pre-training or fine-tuning satellite foundation models on a 600M+ labelled multi-crop agricultural image corpus (the single most valuable asset here, and one of very few of its kind); perennial-crop yield model training with genuine block-level ground truth rather than district statistics; forecast-versus-realised series as backtest targets for soft-commodity supply models (sugar, coffee, palm oil, nuts); benchmark sets for remote-sensing crop segmentation and tree-counting models; hedging and offtake-contract sizing models; harvest labour and logistics optimisation; climate-impact models spanning every major climate zone

The data is with 10 years of history.

Coverage spans Other, US, Europe, APAC; Application Software, Agricultural Products & Services; geospatial, alternative, supply_chain, fundamental; commodities, equities, derivatives.