Compositional Data Package
Freeware, stand-alone, multi-platform software!
Building on the log-ratio framework pioneered by John Aitchison in the early 1980s, statistical analysis of compositional data is not straightforward to perform using ordinary statistical packages. To address this issue, we have developed a free open-source stand-alone software tool, the Compositional Data Package (CoDaPack in short), which implements the basic compositional statistical analyses and graphical representations through an a user-friendly point-and-click interface. The package is designed having in mind practitioners in the applied sciences who typically have little experience with statistical computing and programming. CoDaPack offers a broad range of features, including:
- Data representation via logratios: alr, clr, and olr through sequential binary partition.
- Operations in the simplex, such as centering, perturbation, power transformation, amalgamation, subcomposition (closure), and zero and missing imputation procedures.
- 2-D and 3-D graphical outputs, such as ternary diagrams, alr-plots, clr-plots, compositional biplots.
- Both compositional descriptive statistics and multivariate analyses such as clustering, linear regression and discriminant analysis.
CoDaPack is widely-used by researchers in different fields. As an open software package, it also allows users to modify and distribute it, providing them with the freedom to have it tailored to their specific needs. We welcome feedback and suggestions for future versions of the CoDaPack.