Learning in the greenhouse accelerates adoption of data-driven cultivation

Data and models are widely available in greenhouse horticulture, yet application often lags when the translation to daily practice is missing. The Human Capital Agenda of NXTGEN Hightech Agrifood therefore focuses on learning in real-world settings. During the kick-off of the Fontys research assignment, Applied Mathematics students were introduced to a realistic challenge and saw firsthand in the greenhouse what is measured and why. This helps develop talent that can not only work with data, but also place it within the daily reality of cultivation decisions. That is exactly where the sector is looking to accelerate.

Students model values between sensor locations in a greenhouse using datasets from the Robocrops datahub, the digital field lab containing data from NXTGEN greenhouse horticulture. The assignment is recognisable for professionals, as spatial patterns directly influence irrigation, climate control and crop health. By defining the purpose of the model from the outset, students learn what an outcome should explain in practice and how to responsibly deal with uncertainty.

At Tomatoworld, students explored where sensors are placed, what they measure and for what purpose. Jan Enthoven provided context for the measurements and showed how variation per row and height affects the overall picture. This creates awareness that data quality and measurement setup influence a model just as much as the methodology itself. As a result, deeper analysis becomes both logical and targeted.

Lars van der Lely of 30MHz explained how the Robocrops datahub collects and makes data accessible. A consistent infrastructure reduces noise and enables comparisons across crops and seasons. This is precisely what allows data to support decision-making rather than becoming an end in itself.

At the World Horti Center, Liesbeth Leurs and Daniëlle Jansen shared examples of decisions supported by these types of models. Students discussed which margins of error are acceptable, when additional measurements are needed and how value can be explained to a grower. In this way, theory is given a framework that professionals recognise and can evaluate.