useR! 2024

The study of rotational movement is of paramount importance in robotics as well as in health science. The statistical unit behind measurements of rotational motion is a sequence of 3-dimensional rotations that evolve over time. The goal of {squat} is to provide accessibility to extensions of common statistical methods for the analysis of rotation-valued time series and functional data. The package relies on the Quaternion class from the Eigen library accessed through the {RcppEigen} package. It provides dedicated classes for a single curve as well as a set of curves. Currently, it supports centring, standardisation, visualisation (powered by {ggplot2} and, optionally, {gganimate}), mean and median computation, random sampling, exponential and logarithmic maps to go back and forth from and to the tangent space respectively, smoothing, resampling, distance matrix computation, clustering methods (hierarchical, k-means and dbscan) and principal component analysis. Clustering and PCA also have their dedicated visualisation tools. The package has a dedicated website (https://lmjl-alea.github.io/squat/index.html) as well as a public Github repository (https://github.com/LMJL-Alea/squat/).