Python package to make statistical test and add statistical annotations on plots generated with Plotly
Project description
🚩 Index of Contents
📌 What is TAP?
Python package to make statistical tests and add statistical annotations on plot generated with Plotly
✅ Features
-
Single function to make statistical tests and add statistical annotations on plot generated with Plotly:
- Box plots
- Strip plots
-
Integrated statistical tests (
scipy.stats
methods):- Mann-Whitney test
- t-test (independent and paired)
- t-test-related (dipendent)
- Levene test
- Wilcoxon test
- Kruskal-Wallis test
- Brunner-Munzel test
- Ansari-Bradley test
- CramerVon-Mises test
- Kolmogorov-Smirnov test
- Alexander-Govern test
- Fligner-Killeen test
- Bartlett test
-
Correction for statistical tests can be applied (
statsmodel.stats.multitest.multipletests
method):- Bonferroni
- Sidak
- Holm-Sidak
- Benjamini-Hochberg
-
Exporting plots to formats:
- png
- jpeg
- webp
- svg
- html
📦 Installation
TAP is present on pipy, and can be downloaded directly with pip
pip install taplib
Or if you prefer you can clone the repository and install it manually
git clone https://github.com/FedericaPersiani/tap.git
cd tap
pip install .
🔍 Example
Once your dataframe has been loaded you can pass it to the plot_stats function which will apply the Mann-Whitney test by default on all classes present in the column indicated as x, using the y column as the value
import tap
import seaborn as sns
df = sns.load_dataset("tips")
x = "day"
y = "total_bill"
tap.plot_stats(df, x, y)
Cutoff pvalue: You can change the significance of the null hypothesis through the cutoff_pvalue parameter, by default it is set to 0.05.
tap.plot_stats(df, x, y, cutoff_pvalue=0.01)
Type test: You can change the test type using the type_test parameter
tap.plot_stats(df, x, y, type_test="CramerVon-Mises")
Type correction: You can apply a p-value correction algorithm via the type_correction parameter
tap.plot_stats(df, x, y, type_correction="Bonferroni")
Order: You can change the sorting of the plot by passing the list with all the entries present in the x column ordered as you prefer
tap.plot_stats(df, x, y, order=["Thur", "Fri", "Sat", "Sun"])
Type plot: You can change the plot type using the type_plot parameter
tap.plot_stats(df, x, y, type_plot="strip")
Pairs: You can decide the pairs that will be used to generate the statistics to plot
tap.plot_stats(df, x, y, pairs=[("Sun", "Sat"), ("Sun", "Thur")])
Sub category: Through the subcategory parameter it is possible to divide the various entries into a further sub-category, you can decide the various pairings using the pairs parameter but in this case you will need to declare them as a tuple (primary category, subcategory)
tap.plot_stats(df, x, y, subcategory="sex")
tap.plot_stats(df, x, y, subcategory="sex", pairs=[(("Sun", "Male"), ("Sat", "Male")), (("Sun", "Male"), ("Sun", "Female"))])
Filename: To directly export the image you can use the filename parameter, the standard export size is (800, 600, 3) but you can modify it via the export_size parameter (width, height, scale-factor)
tap.plot_stats(df, x, y, filename="images/export_1.png", export_size=(800, 400, 3))
Kwargs: Through the kwargs parameter you can pass a key/value pairs directly to the plotly function, such as the size of the figure, or a title
tap.plot_stats(df, x, y, kwargs={"width":500, "height":500, "title": "My title"})
📝 Similar work
This repository is inspired by trevismd/statannotations (Statannotations), which compute statistical tests and annotations with seaborn
💬 Citation
BibTeX
@software{persiani_2024_10604555,
author = {Persiani, Federica and
Malori, Damiano},
title = {Discovery-Circle/tap: v0.1.3},
month = feb,
year = 2024,
publisher = {Zenodo},
version = {0.1.3},
doi = {10.5281/zenodo.10604555},
url = {https://doi.org/10.5281/zenodo.10604555}
}
APA
Persiani, F., & Malori, D. (2024). Discovery-Circle/tap: v0.1.3 (0.1.3). Zenodo. https://doi.org/10.5281/zenodo.10604555
✨ Contributors
Federica Persiani 💻 🔬 |
Damiano Malori 💻 📦 |
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.