This project aims to apply machine learning in the context of potato farming. By analyzing many previous spectrometer scans of potato plant leaves, models will be developed that can estimate the amount of nutrients in a potato plant based on a single scan. The models produced by this project will help offer a more scalable and economical solution for precisely monitoring potato crops. This will provide a way for farmers to assess plant health much more quickly in the field, which in turn allows them to use resources such as fertilizer and water in a more targeted and efficient way.