Recently, an R package was developed for categorical functional data analysis (CFDA). This statistical approach extends the usual functional data analysis to temporal categorical data, and as such is particularly relevant for TDS data. CFDA produces a PCA-like map of the sensory evaluations (subject × product) based on the sequences of sensations. Each axis represents leading temporal patterns and the coordinates of sensory evaluations on these axes depict their main temporal characteristics. Then, those coordinates can be used as inputs for further statistical analyses such as clustering of subjects or discriminant analysis of products, both based on temporal perception of the products. Classical analysis of TDS data consists of a series of independent analyses of specific variables: number of citations, dominance rate, duration of dominance or transitions. CFDA presents the advantage of dealing with the entire TDS signal in the same analysis. This paper demonstrates the relevance of CFDA for the analysis of TDS data by using pedagogical data and a real TDS dataset.