Compoda

Compositional data analysis tools implemented in python.

Currently, this library is primarily being developed for (but not limited to) magnetic resonance images with multiple contrasts. For further details, please see my paper here (or here).

Dependencies

Python 3

Package	Tested version
NumPy	1.15.4
Scipy	1.2.0

Additionally required for example scripts:

Package	Tested version
matplotlib	3.0.2
NiBabel	2.2.1

Installation

Run this command in your command line:

pip install compoda

or as an alternative

• Clone this repository and change directory to:

cd /path/to/compoda

• Install the requirements by running the following command:

pip install -r requirements.txt

• Install compoda:

python setup.py install

Support

Please use GitHub issues for questions, bug reports or feature requests.

License

The project is licensed under BSD-3-Clause.

References

• Compositional data analysis in a nutshell.

• Aitchison, J. (1982). The Statistical Analysis of Compositional Data. Journal of the Royal Statistical Society, 44(2), 139–177.

• Aitchison, J. (2002). A Concise Guide to Compositional Data Analysis. CDA Workshop Girona, 24, 73–81.

• Pawlowsky-Glahn, V., Egozcue, J. J., & Tolosana-Delgado, R. (2015). Modelling and Analysis of Compositional Data. Chichester, UK: John Wiley & Sons, Ltd. http://doi.org/10.1002/9781119003144