taplib 0.1.0

Python package to make statistical test and add statistical annotations on plots generated with Plotly

🚩 Index of Contents

• What is TAP?
• Features
• Installation
• Examples
• Similar work
• Contributors
• Citation

📌 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
  • pdf
  • 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_10464613,
  author       = {Persiani, Federica and
                  Malori, Damiano},
  title        = {Discovery-Circle/tap: v0.1.0},
  month        = jan,
  year         = 2024,
  publisher    = {Zenodo},
  version      = {0.1.0},
  doi          = {10.5281/zenodo.10464613},
  url          = {https://doi.org/10.5281/zenodo.10464613}
}

APA

Persiani, F., & Malori, D. (2024). Discovery-Circle/tap: v0.1.0 (0.1.0). Zenodo. https://doi.org/10.5281/zenodo.10464613

✨ Contributors

Federica Persiani 💻 🔬

Damiano Malori 💻 📦