MIDS Capstone Project Spring 2016

Soybean Yield Prediction: Using Satellite Imagery to Predict Commodity Yields

Our project aimed to predict soybean yields using ONLY satellite and weather data. We studied different modelling techniques to determine which are better for this purpose. We also attempted to make predictions earlier in the soybean growth cycle, striving for similar or better accuracy than gold-standard USDA models. Since our cost basis for this prediction is much lower than the USDA's (which utilizes a significantly more hands-on and labor intensive approach), "close enough" modelling results have tremendous advantage, and this information is extremely valuable to commodities markets.

Last updated:

March 30, 2017