Compositional time series: Past and perspectives

This work reviews academic contributions that focus on the analysis of compositional dynamic series, a little investigated topic in spite of the wide data availability in social sciences. It explores the available options and divides research articles into two main probability approaches, frequentist and Bayesian, and enumerates several specific data transformations and techniques. As conclusion, this branch of the compositional statistical analysis requires a profound updating and, for this very same reason, is a fertile field for future research.