Grape maturity predictions are based on a maturation model built up over the data available in the pool. There are two models offered: generic and tailored.
As part of the individual subscription, we use a generic maturation model based on ALL the data for the varieties and regions of interest, and as new data (new blocks, new varieties, new regions) is added into the pool this “national” maturation model is refreshed and enhanced.
For enterprise subscriptions, we can also create a maturation model tailored to the company’s varieties and blocks of interest. This results in a model that is more specific to the customer, rather than “averaged” with all other data. Two key points: this data is still added to the pool, and forms part of the input to the generic model; and the tailored and generic models can be used for a comparison, to highlight where an enterprise customer’s blocks differ from an industry-wide model.
As data accumulates in the pool, all models and predictions will become more complete and robust – that is the main purpose of the accumulation exercise. In addition to that, as the data pool grows in size and scope, we will look for more predictive elements, and other models that can be created from it. One possibility is an improved yield estimation; another might be the addition of soil characteristics or water usage into the maturity prediction.