Abstract
Between 2012 and 2014, 715 green coffee samples were gathered by Almacafé S.A. (Bogotá, Colombia) from 27 countries. These were analysed at the nuclear magnetic resonance (NMR) laboratory at Universidad del Valle (Cali, Colombia). Over 1000 methanolic coffee extracts were prepared and 4563 spectra were acquired in a fully automatic manner using a 400 MHz NMR spectrometer (Bruker Biospin, Germany). The dataset spans the variance that could be expected for an industrial application of origin monitoring, including samples from different harvest times, collected over several years, and processed by at least two distinct operators. The resulting 1D and 2D spectra can be used to develop and evaluate feature extraction methods, multivariate algorithms, and automation monitoring techniques. They can also be used as datasets for teaching, or as a reference for new studies of similar samples and approaches.