Statistical package in Python based on Pandas
.. -- mode: rst --
|
.. image:: https://badge.fury.io/py/pingouin.svg :target: https://badge.fury.io/py/pingouin
.. image:: https://img.shields.io/conda/vn/conda-forge/pingouin.svg :target: https://anaconda.org/conda-forge/pingouin
.. image:: https://img.shields.io/github/license/raphaelvallat/pingouin.svg :target: https://github.com/raphaelvallat/pingouin/blob/master/LICENSE
.. image:: https://github.com/raphaelvallat/pingouin/actions/workflows/pytest.yml/badge.svg :target: https://github.com/raphaelvallat/pingouin/actions
.. image:: https://codecov.io/gh/raphaelvallat/pingouin/branch/master/graph/badge.svg :target: https://codecov.io/gh/raphaelvallat/pingouin
.. image:: https://pepy.tech/badge/pingouin/month :target: https://pepy.tech/badge/pingouin/month
.. image:: http://joss.theoj.org/papers/d2254e6d8e8478da192148e4cfbe4244/status.svg :target: http://joss.theoj.org/papers/d2254e6d8e8478da192148e4cfbe4244
.. image:: https://pingouin-stats.org/images/logopingouin.png :align: center
Pingouin is an open-source statistical package written in Python 3 and based mostly on Pandas and NumPy. Some of its main features are listed below. For a full list of available functions, please refer to the API documentation <https://pingouin-stats.org/api.html>_.
- ANOVAs: N-ways, repeated measures, mixed, ancova
- Pairwise post-hocs tests (parametric and non-parametric) and pairwise correlations
- Robust, partial, distance and repeated measures correlations
- Linear/logistic regression and mediation analysis
- Bayes Factors
- Multivariate tests
- Reliability and consistency
- Effect sizes and power analysis
- Parametric/bootstrapped confidence intervals around an effect size or a correlation coefficient
- Circular statistics
- Chi-squared tests
- Plotting: Bland-Altman plot, Q-Q plot, paired plot, robust correlation...
For example, the :code:ttest_ind function of SciPy returns only the T-value and the p-value. By contrast, the :code:ttest function of Pingouin returns the T-value, the p-value, the degrees of freedom, the effect size (Cohen's d), the 95% confidence intervals of the difference in means, the statistical power and the Bayes Factor (BF10) of the test.
Documentation =============
Link to documentation <https://pingouin-stats.org/index.html>_
If you have questions, please ask them in GitHub Discussions <https://github.com/raphaelvallat/pingouin/discussions>_.
Installation ============
Dependencies
The main dependencies of Pingouin are:
NumPy <https://numpy.org/>_ >= 1.22.4SciPy <https://www.scipy.org/>_ >= 1.8.0Pandas <https://pandas.pydata.org/>_ >= 2.1.1Pandas-flavor <https://github.com/Zsailer/pandasflavor>Statsmodels <https://www.statsmodels.org/>_ >= 0.14.1Matplotlib <https://matplotlib.org/>_Seaborn <https://seaborn.pydata.org/>_Scikit-learn <https://scikit-learn.org/>_ >= 1.2.2Tabulate <https://github.com/astanin/python-tabulate>_
Mpmath <http://mpmath.org/>_
User installation
Pingouin can be easily installed using uv <https://docs.astral.sh/uv/>_
.. code-block:: shell
uv pip install pingouin
pip
.. code-block:: shell
pip install pingouin
or conda
.. code-block:: shell
conda install -c conda-forge pingouin
New releases are frequent so always make sure that you have the latest version:
.. code-block:: shell
uv pip install --upgrade pingouin
Development
To build and install from source, clone this repository and install in editable mode with uv <https://docs.astral.sh/uv/>_
.. code-block:: shell
git clone https://github.com/raphaelvallat/pingouin.git cd pingouin uv pip install --group=dev --editable .
# test the package pytest --verbose
Quick start ============
Click on the link below and navigate to the notebooks/ folder to run a collection of interactive Jupyter notebooks showing the main functionalities of Pingouin. No need to install Pingouin beforehand, the notebooks run in a Binder environment.
.. image:: https://mybinder.org/badge.svg :target: https://mybinder.org/v2/gh/raphaelvallat/pingouin/develop
10 minutes to Pingouin
- T-test
.. code-block:: python
import numpy as np import pingouin as pg
np.random.seed(123) mean, cov, n = [4, 5], [(1, .6), (.6, 1)], 30 x, y = np.random.multivariate_normal(mean, cov, n).T
# T-test pg.ttest(x, y)
.. table:: Output :widths: auto
====== ===== ============= ======= ============= ========= ====== ======= T dof alternative pval CI95 cohend BF10 power ====== ===== ============= ======= ============= ========= ====== ======= -3.401 58 two-sided 0.001 [-1.68 -0.43] 0.878 26.155 0.917 ====== ===== ============= ======= ============= ========= ====== =======
- Pearson's correlation
.. code-block:: python
pg.corr(x, y)
.. table:: Output :widths: auto
=== ===== =========== ======= ====== ======= n r CI95 p_val BF10 power === ===== =========== ======= ====== ======= 30 0.595 [0.3 0.79] 0.001 69.723 0.950 === ===== =========== ======= ====== =======
- Robust correlation
.. code-block:: python
# Introduce an outlier x[5] = 18 # Use the robust biweight midcorrelation pg.corr(x, y, method="bicor")
.. table:: Output :widths: auto
=== ===== =========== ======= ======= n r CI95 p_val power === ===== =========== ======= ======= 30 0.576 [0.27 0.78] 0.001 0.933 === ===== =========== ======= =======
- Test the normality of the data
The pingouin.normality function works with lists, arrays, or pandas DataFrame in wide or long-format.
.. code-block:: python
print(pg.normality(x)) # Univariate normality print(pg.multivariatenormality(np.columnstack((x, y)))) # Multivariate normality
.. table:: Output :widths: auto
===== ====== ======== W pval normal ===== ====== ======== 0.615 0.000 False ===== ====== ========
.. parsed-literal::
(False, 0.00018)
- One-way ANOVA using a pandas DataFrame
.. code-block:: python
# Read an example dataset df = pg.readdataset('mixedanova')
# Run the ANOVA aov = pg.anova(data=df, dv='Scores', between='Group', detailed=True) print(aov)
.. table:: Output :widths: auto
======== ======= ==== ===== ======= ======= ======= Source SS DF MS F p_unc np2 ======== ======= ==== ===== ======= ======= ======= Group 5.460 1 5.460 5.244 0.023 0.029 Within 185.343 178 1.041 nan nan nan ======== ======= ==== ===== ======= ======= =======
- Repeated measures ANOVA
.. code-block:: python
pg.rm_anova(data=df, dv='Scores', within='Time', subject='Subject', detailed=True)
.. table:: Output :widths: auto
======== ======= ==== ===== ======= ======= ======= ======= Source SS DF MS F p_unc ng2 eps ======== ======= ==== ===== ======= ======= ======= ======= Time 7.628 2 3.814 3.913 0.023 0.04 0.999 Error 115.027 118 0.975 nan nan nan nan ======== ======= ==== ===== ======= ======= ======= =======
- Post-hoc tests corrected for multiple-comparisons
.. code-block:: python
# FDR-corrected post hocs with Hedges'g effect size posthoc = pg.pairwise_tests(data=df, dv='Scores', within='Time', subject='Subject', parametric=True, padjust='fdr_bh', effsize='hedges')
# Pretty printing of table pg.print_table(posthoc, floatfmt='.3f')
.. table:: Output :widths: auto
========== ======= ======= ======== ============ ====== ====== ============= ======= ======== ========== ====== ======== Contrast A B Paired Parametric T dof alternative punc pcorr p_adjust BF10 hedges ========== ======= ======= ======== ============ ====== ====== ============= ======= ======== ========== ====== ======== Time August January True True -1.740 59.000 two-sided 0.087 0.131 fdr_bh 0.582 -0.328 Time August June True True -2.743 59.000 two-sided 0.008 0.024 fdr_bh 4.232 -0.483 Time January June True True -1.024 59.000 two-sided 0.310 0.310 fdr_bh 0.232 -0.170 ========== ======= ======= ======== ============ ====== ====== ============= ======= ======== ========== ====== ========
- Two-way mixed ANOVA
.. code-block:: python
# Compute the two-way mixed ANOVA aov = pg.mixed_anova(data=df, dv='Scores', between='Group', within='Time', subject='Subject', correction=False, effsize="np2") pg.print_table(aov)
.. table:: Output :widths: auto
=========== ===== ===== ===== ===== ===== ======= ===== ======= Source SS DF1 DF2 MS F p_unc np2 eps =========== ===== ===== ===== ===== ===== ======= ===== ======= Group 5.460 1 58 5.460 5.052 0.028 0.080 nan Time 7.628 2 116 3.814 4.027 0.020 0.065 0.999 Interaction 5.167 2 116 2.584 2.728 0.070 0.045 nan =========== ===== ===== ===== ===== ===== ======= ===== =======
- Pairwise correlations between columns of a dataframe
.. code-block:: python
import pandas as pd np.random.seed(123) z = np.random.normal(5, 1, 30) data = pd.DataFrame({'X': x, 'Y': y, 'Z': z}) pg.pairwise_corr(data, columns=['X', 'Y', 'Z'], method='pearson')
.. table:: Output :widths: auto
=== === ======== ============= === ===== ============= ======= ====== ======= X Y method alternative n r CI95 p_unc BF10 power === === ======== ============= === ===== ============= ======= ====== ======= X Y pearson two-sided 30 0.366 [0.01 0.64] 0.047 1.500 0.525 X Z pearson two-sided 30 0.251 [-0.12 0.56] 0.181 0.534 0.272 Y Z pearson two-sided 30 0.020 [-0.34 0.38] 0.916 0.228 0.051 === === ======== ============= === ===== ============= ======= ====== =======
- Pairwise T-test between columns of a dataframe
.. code-block:: python
data.ptests(paired=True, stars=False)
.. table:: Pairwise T-tests, with T-values on the lower triangle and p-values on the upper triangle :widths: auto
==== ====== ====== ===== .. X Y Z ==== ====== ====== ===== X - 0.226 0.165 Y -1.238 - 0.658 Z -1.424 -0.447 - ==== ====== ====== =====
- Multiple linear regression
.. code-block:: python
pg.linear_regression(data[['X', 'Z']], data['Y'])
.. table:: Linear regression summary :widths: auto
========= ====== ===== ====== ====== ===== ======== ========== =========== names coef se T pval r2 adj_r2 CI2.5 CI97.5 ========= ====== ===== ====== ====== ===== ======== ========== =========== Intercept 4.650 0.841 5.530 0.000 0.139 0.076 2.925 6.376 X 0.143 0.068 2.089 0.046 0.139 0.076 0.003 0.283 Z -0.069 0.167 -0.416 0.681 0.139 0.076 -0.412 0.273 ========= ====== ===== ====== ====== ===== ======== ========== ===========
- Mediation analysis
.. code-block:: python
pg.mediationanalysis(data=data, x='X', m='Z', y='Y', seed=42, nboot=1000)
.. table:: Mediation summary :widths: auto
======== ====== ===== ====== ========== =========== ===== path coef se pval CI2.5 CI97.5 sig ======== ====== ===== ====== ========== =========== ===== Z ~ X 0.103 0.075 0.181 -0.051 0.256 No Y ~ Z 0.018 0.171 0.916 -0.332 0.369 No Total 0.136 0.065 0.047 0.002 0.269 Yes Direct 0.143 0.068 0.046 0.003 0.283 Yes Indirect -0.007 0.025 0.898 -0.069 0.029 No ======== ====== ===== ====== ========== =========== =====
- Contingency analysis
.. code-block:: python
data = pg.readdataset('chi2independence') expected, observed, stats = pg.chi2_independence(data, x='sex', y='target') stats
.. table:: Chi-squared tests summary :widths: auto
================== ======== ====== ===== ===== ======== ======= test lambda chi2 dof p cramer power ================== ======== ====== ===== ===== ======== ======= pearson 1.000 22.717 1.000 0.000 0.274 0.997 cressie-read 0.667 22.931 1.000 0.000 0.275 0.998 log-likelihood 0.000 23.557 1.000 0.000 0.279 0.998 freeman-tukey -0.500 24.220 1.000 0.000 0.283 0.998 mod-log-likelihood -1.000 25.071 1.000 0.000 0.288 0.999 neyman -2.000 27.458 1.000 0.000 0.301 0.999 ================== ======== ====== ===== ===== ======== =======
Integration with Pandas
Several functions of Pingouin can be used directly as pandas DataFrame methods. Try for yourself with the code below:
.. code-block:: python
import pingouin as pg
# Example 1 | ANOVA df = pg.readdataset('mixedanova') df.anova(dv='Scores', between='Group', detailed=True)
# Example 2 | Pairwise correlations data = pg.read_dataset('mediation') data.pairwise_corr(columns=['X', 'M', 'Y'], covar=['Mbin'])
# Example 3 | Partial correlation matrix data.pcorr()
The functions that are currently supported as pandas method are:
pingouin.anova <https://pingouin-stats.org/generated/pingouin.anova.html#pingouin.anova>_pingouin.ancova <https://pingouin-stats.org/generated/pingouin.ancova.html#pingouin.ancova>_pingouin.rmanova <https://pingouin-stats.org/generated/pingouin.rmanova.html#pingouin.rmanova>pingouin.mixedanova <https://pingouin-stats.org/generated/pingouin.mixedanova.html#pingouin.mixedanova>pingouin.welchanova <https://pingouin-stats.org/generated/pingouin.welchanova.html#pingouin.welchanova>pingouin.pairwisetests <https://pingouin-stats.org/generated/pingouin.pairwisetests.html#pingouin.pairwisetests>pingouin.pairwisetukey <https://pingouin-stats.org/generated/pingouin.pairwisetukey.html#pingouin.pairwisetukey>pingouin.pairwisecorr <https://pingouin-stats.org/generated/pingouin.pairwisecorr.html#pingouin.pairwisecorr>pingouin.partialcorr <https://pingouin-stats.org/generated/pingouin.partialcorr.html#pingouin.partialcorr>pingouin.pcorr <https://pingouin-stats.org/generated/pingouin.pcorr.html#pingouin.pcorr>_pingouin.rcorr <https://pingouin-stats.org/generated/pingouin.rcorr.html#pingouin.rcorr>_pingouin.ptests <https://pingouin-stats.org/generated/pingouin.ptests.html#pingouin.ptests>_pingouin.mediationanalysis <https://pingouin-stats.org/generated/pingouin.mediationanalysis.html#pingouin.mediationanalysis>
Pingouin was created and is maintained by Raphael Vallat <https://raphaelvallat.github.io>_, a postdoctoral researcher at UC Berkeley, mostly during his spare time. Contributions are more than welcome so feel free to contact me, open an issue or submit a pull request!
To see the code or report a bug, please visit the GitHub repository <https://github.com/raphaelvallat/pingouin>_.
This program is provided with NO WARRANTY OF ANY KIND. Pingouin is still under heavy development and there are likely hidden bugs. Always double check the results with another statistical software.
Contributors
- Nicolas Legrand
Richard Höchenberger <http://hoechenberger.net/>_Arthur Paulino <https://github.com/arthurpaulino>_Eelke Spaak <https://eelkespaak.nl/>_Johannes Elfner <https://www.linkedin.com/in/johannes-elfner/>_Stefan Appelhoff <https://stefanappelhoff.com>_
If you want to cite Pingouin, please use the publication in JOSS:
Vallat, R. (2018). Pingouin: statistics in Python. Journal of Open Source Software*, 3(31), 1026, https://doi.org/10.21105/joss.01026 <https://doi.org/10.21105/joss.01026>_
Acknowledgement ===============
Several functions of Pingouin were inspired from R or Matlab toolboxes, including:
effsize package (R) <https://cran.r-project.org/web/packages/effsize/effsize.pdf>_ezANOVA package (R) <https://cran.r-project.org/web/packages/ez/ez.pdf>_pwr package (R) <https://cran.r-project.org/web/packages/pwr/pwr.pdf>_circular statistics (Matlab) <https://www.mathworks.com/matlabcentral/fileexchange/10676-circular-statistics-toolbox-directional-statistics>_robust correlations (Matlab) <https://sourceforge.net/projects/robustcorrtool/>_repeated-measure correlation (R) <https://cran.r-project.org/web/packages/rmcorr/index.html>_real-statistics.com <https://www.real-statistics.com/>_
