pyCoDaMath 1.0

Compositional data (CoDa) analysis tools for Python

pyCoDaMath

pyCoDaMath provides compositional data (CoDa) analysis tools for Python

• Source code: https://bitbucket.org/genomicepidemiology/pycoda

Getting Started

This package extends the Pandas dataframe object with various CoDa tools. It also provides a set of plotting functions for CoDa figures.

Installation

Clone the git repo to your local hard drive:

git clone https://[e-mail]/genomicepidemiology/pycoda.git

Enter pycoda directory and type

pip install ./

Usage

The pyCoDaMath module is loaded as

import pycodamath

At this point, in order to get CLR values from a Pandas DataFrame df, do

df.coda.clr()

Documentation

CLR transformation - point estimate

df.coda.clr()

Returns centered logratio coefficients. If the data frame contains zeros, values will be replaced by the Aitchison mean point estimate.

CLR transformation - standard deviation

df.coda.clr_std(n_samples=5000)

Returns the standard deviation of n_samples random draws in CLR space.

Parameters

• n_samples (int) - Number of random draws from a Dirichlet distribution.

ALR transformation - point estimate

df.coda.alr(part=None)

Same as clr() but returning additive logratio values. If part is None, then the last part of the composition is used, otherwise part is used as denominator.

Parameters

• part (str) - Name of the part to be used as denominator.

ALR transformation - standard deviation

df.coda.alr_std(part=None, n_samples=5000)

Same as clr_std, but in ALR space.

Parameters

• part (str) - Name of the part to be used as denominator.

• n_samples (int) - Number of random draws from a Dirichlet distribution.

ILR transformation - point estimate

df.coda.ilr(psi=None)

Same as clr() but for isometric logratio transform. An orthonormal basis can be provided as psi. If no basis is given, a default sequential binary partition basis will be used.

Parameters

• psi (array_like) - Orthonormal basis.

ILR transformation - standard deviation

df.coda.ilr_std(psi=None, n_samples=5000)

This method does not exist (yet).

Bayesian zero replacement

df.coda.zero_replacement(n_samples=5000)

Returns a count table with zero values replaced by finite values using Bayesian inference.

Parameters

• n_samples (int) - Number of random draws from a Dirichlet distribution.

Closure

df.coda.closure(N)

Apply closure to constant N to the composition.

Parameters

• N (int) - Closure constant.

Total variance

df.coda.totvar()

Calculates the total variance of a set of compositions.

Geometric mean

df.coda.gmean()

Calculates the geometric mean of a set of compositions.

Centering

df.coda.center()

Centers (and scales) the composition by dividing by the geometric mean and powering by the reciprocal variance.

Plotting functions

PCA biplot

class pycoda.pca.Biplot(data, default=True)

Plots a PCA biplot. Set default to False for an empty plot. The parameter data (DataFrame) is the data to be analyzed. Use counts, not CLR values.

A number of methods are available for customizing the biplot:

• plotloadings(cutoff=0, scale=None, labels=None)
• plotloadinglabels(labels=None)
• plotscores(group=None, palette=None, legend=True, labels=None)
• plotscorelables(labels=None)
• plotellipses(group=None, palette=None)
• plotcentroids(group=None, palette=None)
• plothulls(group=None, palette=None)
• plotcontours(group=None, palette=None, size=None, levels=None)
• removepatches()
• removescores()
• removelabels()

The keyword labels is a list of labelnames. If labels is None, all labels are plottet. Use labels=[] for no labels.

The keyword group is a Pandas dataframe with index equal to the index of data.

The keyword palette is a dict with colors to use to each unique member of group.

Example import pycoda as coda import pandas as pd

data = pd.read_csv('example/kilauea_iki_chem.csv')
mypca = coda.pca.Biplot(data)
mypca.plothulls()
mypca.removelabels()
mypca.plotloadinglabels(['FeO'])

Ternary diagram

pycoda.plot.ternary()