Fourier Transform Near Infrared (FT-NIR) spectra and sensory scores in green and roasted specialty coffee for machine learning-based quality monitoring

This dataset provides raw and pre-processed Fourier Transform Near Infrared (FT-NIR) spectra for green and roasted specialty coffee samples, along with their sensory quality scores. Pre-processing techniques include baseline correction, area normalization, Multiplicative Scatter Correction (MSC), and first and second derivatives. Additionally, the dataset includes computational tools (programmed in R statistical software) for spectral pre-processing and calibration of machine learning models designed to predict sensory quality. These R-codes enable the development of predictive tools for non-destructive, real-time quality assessment, facilitating sensory analysis and supporting quality monitoring processes in the specialty coffee industry.