Yield measurements in aeroponics for six different crops

Our experiments, conducted during production stages, included six different crops: garlic chives, basil, red chard, rainbow chard, arugula and mint. It contains hourly values of room carbon-di-oxide (CO2) levels [ppm], room relative humidity [%], room light level [lux], room temperature [°F], room vapour pressure deficit (VPD) [mbar], reservoir water pH, water total dissolved solids (TDS) [ppm] and reservoir temperature [°F]. Manually collected data include number of days in tray, number of days in tower, harvest number (how many times the plant has been harvested since first transplanted to AeroPod) and grow number (how many times a plant has been transplanted into that spot). The target value that the AI models use is the crop yield (oz/spot then converted to g/spot). For this study, 200 samples have been collected between November 2018 and August 2019. The data was used to construct two different datasets. The first is a completely tabular dataset in which the sensors readings were averaged. The averages were calculated in such a way that each sensor contributed with one scalar feature per data point. The second is a mixed dataset in which the tabular data were kept unchanged, and in which each sensor contributed with a vector containing the time series response per data point.