Main content area

Coffee terroir: cupping description profiles and their impact upon prices in Central American coffees

Conley, Jamison, Wilson, Bradley
GeoJournal 2020 v.85 no.1 pp. 67-79
artificial intelligence, auctions, crop production, flavor, growers, prices, taste, Central America
We examine the gustatory properties and flavors of 742 coffees from the Cup of Excellence program in Central America to determine if there are distinct quality profiles for each of five coffee-growing countries, which can provide evidence of coffee terroir. We further compare these properties with the price of a coffee at auction to answer the following research questions: (1) Does the presence or absence of specific properties impact a coffee’s price at auction? (2) Can flavor profiles be used to identify a coffee’s country of origin? (3) Are the effects of descriptors upon price mitigated by country, or do they vary by country? Employing statistical and machine learning techniques, we find evidence of terroir in coffee production, and that the country-scale flavor profiles resulting from this can have a substantial impact on the coffee’s auction price. This information can potentially assist coffee growers in aligning their product’s qualities with the qualities their buyers are seeking. It can also enable the development of formal appellations to confirm each country’s unique coffee profile.