Until now green fingers were enough to make decisions or to make predictions, but we do see that this is becoming more difficult. There is a need for more accurate predictions and the number of people with green fingers is decreasing dramatically. Luckily Artificial Intelligence (AI) can help in the horticultural sector to compensate for the loss of knowledge. The LetsGrow.com yield prediction is the first step towards digitizing tomato production.
The data that LetsGrow.com transfers to accurate and trustworthy prognoses contribute towards reaching a better results for the customers. Growers(associations) and trading companies are struggling with this for several years already. “How much production can we expect this period?”, “What is my expected labor demand?”. The figures of the manual prognoses differ often by a double-digit percentage from what is actually produced. LetsGrow.com has with the help of machine learning a model which makes it possible to predict yield with an accuracy between 83% and 93%. This is an enormous leap forward.
In 2018 LetsGrow.com launched its yield prediction. This prediction was received very positively by its customers. In Q4 of 2017, LetsGrow.com did research on the possibilities to use Machine Learning in creating yield predictions. Since its release, roughly ten large tomato growers in the Netherlands have successfully utilized the yield prediction module.
The methodology used to achieve accurate and reliable yield predictions is based on data from several years of production from Greenco. Greenco is an international grower and trader of the best snack vegetables: snacktomoto’s, snack peppers, and snack cucumbers. Through the usage of machine learning to get the right correlation between the current production and the past production the production for the coming weeks can be predicted. Every year more data is entered into the prediction model making it even more accurate.
For years already LetsGrow.com is collecting data from and for customers. How nice would it be if all this data can be used for each individual customer to make an accurate prognoses model? By combining our background in plant physiology with machine learning we can easily switch to make very accurate analyses like the yield prediction for customers.
Supply and demand
In the greenhouse supply chain, there is a growing demand for better matching supply and demand. Because of a mismatch in supply and demand, there is a major distortion in the market. Going from oversupply to shortage. Resulting in pressure on the pricing of the product. With a good prediction supply and demand can find a better balance resulting in better sales prices. There would be less oversupplying of products in the market.