josephlalli / statannotations

add statistical annotations (pvalue significance) on existing boxplot/barplot generated by seaborn

What is it

Python package to optionnally compute statistical test and add statistical annotations on an existing boxplot/barplot generated by seaborn.

Derived work

This repositoty is evolving independently from webermarcolivier/statannot by Marc Weber. It is based on commit 1835078 of Feb 21, 2020, tagged "v0.2.3".

Additions/modifications since that fork are below represented in bold (previous fixes are not listed). New issues and PRs are welcome and will be looked into.

Contrary to statannot, this repository does not publish releases on pypi.org. The statannot interface, at least until its version 0.2.3, is directly usable in statannotations, which provides additional features.

Features

• Single function to add statistical annotations on an existing boxplot/barplot generated by seaborn boxplot.
• Integrated statistical tests (binding to scipy.stats methods):
  • Mann-Whitney
  • t-test (independent and paired)
  • Welch's t-test
  • Levene test
  • Wilcoxon test
  • Kruskal-Wallis test
• Smart layout of multiple annotations with correct y offsets.
• Annotations can be located inside or outside the plot.
• Corrections for multiple testing can be applied (binding to statsmodels.stats.multitest.multipletests methods):
  • Bonferroni
  • Holm-Bonferroni
  • Benjamini-Hochberg
  • Benjamini-Yekutieli
  • And any other function from any source with minimal extra code
• Format of the statistical test annotation can be customized: star annotation, simplified p-value, or explicit p-value.
• Optionally, custom p-values can be given as input. In this case, no statistical test is performed, but corrections for multiple testing can be applied.

Installation

pip install git+https://github.com/trevismd/statannotations.git

or, from the directory into which this repository was cloned:

pip install -r requirements.txt .

Documentation

See example jupyter notebook doc/example.ipynb.

Usage

Here is a minimal example:

import seaborn as sns
from statannotations import add_stat_annotation

df = sns.load_dataset("tips")
x = "day"
y = "total_bill"
order = ['Sun', 'Thur', 'Fri', 'Sat']
ax = sns.boxplot(data=df, x=x, y=y, order=order)
test_results = add_stat_annotation(
    ax, data=df, x=x, y=y, order=order,
    box_pairs=[("Thur", "Fri"), ("Thur", "Sat"), ("Fri", "Sun")],
    test='Mann-Whitney', text_format='star', loc='outside', verbose=2)

test_results

Examples

Requirements

• Python >= 3.5
• numpy >= 1.12.1
• seaborn >= 0.8.1
• matplotlib >= 2.2.2
• pandas >= 0.23.0
• scipy >= 1.1.0
• statsmodels (optional, for multiple testing corrections)