Combining agronomy and machine learning for accurate yield forecasting

One company is using advanced satellite imaging and AI to give farmers real-time yield predictions

Spotted: Traditionally, specialist crop farmers rely on dated yield predictors that, unfortunately, often provide inaccurate or unreliable data. As a result, preparations and infrastructure are not optimised, leading to wasted products and inefficiency in the supply chain. This is where startup Bountiful comes in. 

Bountiful utilises advanced satellite imagery, data science, and expert analysis to provide more accurate yield predictions to farmers, made accessible in the easy-to-understand platform. Taking into account historical data, hyper-localised weather modelling, and crop classification, Bountiful’s proprietary technology and algorithms can help to reduce the common inaccuracies within forecasting by taking more data points into consideration. 

For instance, Bountiful highlights that when the company began in 2015, predictions of grape yield were incorrect 30 per cent of the time. When you contrast this with the 98 per cent accuracy in yield forecasting Bountiful achieved for speciality crops in 2019, the benefit is evident. The platform enables farmers to make month-by-month personalised decisions to benefit their crop and yield, with the predictive algorithms also used to optimise water use and nutrient application. 

Bountiful’s platform is available on a subscription basis to farmers and other players in the speciality crop industry. The forecasting technology is currently being used to predict grape and almond yields across California and Australia, but the company is looking to add walnuts and pistachios to its repertoire and customers can also request new locations or crops.  

Springwise has also spotted one company that uses automated robots and AI to aid in regenerative farming as well as a startup using satellites to predict agricultural risk and other challenges in sustainability.

Written By: Archie Cox