Sensory analysis of cafes assumes that a sensory panel is formed by trained panelists according to recommendations of the American Specialty Coffee Association. However, the choice that determines the preference of a coffee is routinely done through experimentation with consumers, in which largely presents no particular skill in terms of sensory characteristics. Upon this fact, this study aimed to conduct a study considering several probabilistic distributions belonging to the class of generalized extreme value, considering a sensory analysis applied to evaluation of four specialty coffees produced with different processes and at different altitudes in the mountain region of the Mantiqueira state of Minas Gerais. For this analysis, we considered a sensory panel trained to untrained consumers. It was found that the extreme value distribution was the best fit and the final note that the odds of a consumer to submit a maximum score was 9.0 points lower. Therefore, there is evidence to conclude that an efficient identification of specialty coffees produced in this region made by consumers requires more intensive training.

In a granulometric analysis of coffee beans with different categories of defects, the data can be organized in contingency tables, and when considering the discrimination by harvest, they may have a structure that suggest a more complex model, by means of the counting of defective coffee beans compared to different crops interacting with the classification of defects and percentages of sieve grains, which characterizes a block design with multivariate responses. However, due to the techniques based on the analysis of variance, considering the uniform correlation structure for all plots, it becomes feasible to propose a model that allows contemplating different structures between the plots, associating the effects of the crops to the defects in the granulometric procedure applied to the coffee beans. Thus, the hypothesis of incorporating the effects of crops associated with defects arises using the biplot multivariate technique. This work aims to propose the use of corrected biplots by predictions obtained trhough the fit to the Generalized Linear Model in the coffee grain size classification, broken down by components of the effect of the harvests. In conclusion, the use of GEE models with the corrected biplot technique by the predictions is feasible for application to be applied to the granulometric analysis of defective coffee beans, presenting discrimination regarding the effects of harvests.