Analyzing compositional data with R

Viac o knihe

This book focuses on the statistical analysis of compositional data sets, such as percentages and proportions. It covers foundational principles and practical applications in descriptive exploratory analysis, robust linear models, and advanced multivariate statistical methods, addressing zeros and missing values while emphasizing data visualization and model display. Illustrated examples and code snippets guide readers in modeling and interpretation. Although primarily a reference for the R package “compositions,” it also serves as a general introduction to Compositional Data Analysis. Awareness of the unique characteristics of compositional data emerged in the Geosciences in the early sixties, but effective strategies were not established until Aitchison's work in the eighties. Since then, research has enhanced our understanding of theoretical principles, potentials, and limitations of interpretation. This comprehensive textbook addresses these issues and their practical implications in software. It is aimed at scientists interested in analyzing compositional data, with applications spanning geosciences, environmental sciences, medicine, official statistics, and economics. Familiarity with basic univariate and multivariate statistics is expected, while knowledge of R is recommended but not mandatory, as the book is self-contained.