Predictive Frost Damage Detection

Frost damage detection in California vineyards using measured chlorophyl values and K-Means clustering statistics to create a new Damage Assessment product offering that provides clients with actionable tools to improve crop yield. Collected GPS coordinates of a known damaged field, converting points to polygon shapefiles to serve as an accuracy assessment of the predictive model using SKlearn confusion matrices. Created a Damage Assessment Map Product that predicts plant damage with ~92% accuracy.

Cross-functional project conducted with Sales, Science and Engineering departments.