Every year, the United States Department of Agriculture (USDA) spends millions of dollars predicting national crop production throughout the growing season. Helped by thousands, they sample farms, by hand, in a portion of the 3,242 counties. Their monthly reports move commodity crop markets—having this information before others provide a significant advantage.
In 2015, Descartes Labs used public satellites, weather data, and their expertise in distributed supercomputing to make the first entirely automated forecast of national corn production. While the USDA reports at a state and national granularity, Descartes Labs can provide per county predictions. Their prediction came within 1.9% of the actual production, reported by the USDA more than 6 months later.
This poster was made to better understand key aspects of corn growth and prediction. It visualizes the 2015 growing season in Polk County, Iowa, and is one of a series. New posters, showing conditions for other counties, can be produced programmatically.