CATPCA in R: Application to Sensorial Data of Wine-Tasting

Analyzing large volumes of categorical data presents a recurring challenge in the social, behavioral, and biological sciences. Therefore, there is a necessity to reduce these data while maintaining minimal loss of information. CATPCA serves as an alternative to conventional Principal Component Analysis (PCA), enabling the reduction of dimensionality for variables measured on different scales. To elucidate the underlying principles of this statistical method, a review of bibliographic sources was undertaken. This search also highlighted the Gifi and the Homals packages, identified as the sole packages equipped with functions supporting the implementation of CATPCA. Both packages were explored to assess and compare their functionalities and features. The Gifi package is derived from the Homals package, aiming for enhanced user-friendliness and flexibility. To further delve into the functionalities and constraints of the Gifi package and to demonstrate the handling of biological data, an analysis of sensory data from wine-tasting trials was additionally conducted. In order to enable users to perform an intuitive and simplified CATPCA, a web application named Gify, available at https://andreiagomes.shinyapps.io/Gify/ , based on the princals() function from the Gifi package, has been developed.