Details
Vinsight forecasts grapes in California. We collect four main data sources that include weather, satellite imagery, crop-specific data and geospatial information as inputs to our models. Machine learning is the foundation for the models as it finds trends to learn patterns in yield variations. Our models ingest snapshots of the remotely sensed data throughout a growing season to forecast yield for that year. Our forecasts are produced at varying spatial levels; state, county and sub-county. Accuracies in the final numbers vary based on spatial scale, our grape forecasts are 90-94% accurate