This talk is about categorical functional data viewed as sample paths of a continuous-time jump process with finite states observed on [0; T]. Principal components and optimal encodings of states are defined in order to represent, visualise and interpret the dynamic of such data. This work is based on the pioneer works of Deville and Saporta in the early 1980's and the paper "Categorical Functional Data Analysis. The cfda R Package. Mathematics 2021, 9, 3074 (https://doi.org/10.3390/math9233074)" ; We illustrate the methodology and the applications using a R shiny interface of the cfda R package with real data.
