boxTestpy

boxTestpy is a beginner-friendly Python package for creating side-by-side boxplots and performing basic two-sample statistical tests (t-test or Mann–Whitney U test) with publication-ready plots. It simplifies data visualization and statistical analysis for beginners and researchers.

Features

• Side-by-side boxplots for a two-level categorical variable
• Normality check using Shapiro–Wilk test
• Automatic selection of independent t-test or Mann–Whitney U test
• Pastel color palette for clean plots
• Optional statistical annotation on plots
• Works with pandas DataFrames

Installation

pip install boxTestpy

For plotting support:

pip install matplotlib seaborn pandas

Quick Start

import pandas as pd
import matplotlib.pyplot as plt
import boxtestpy as btp

# Public health dataset: Blood Pressure (systolic)
data = {
    "group": ["Control"]*10 + ["Treatment"]*10,
    "blood_pressure": [
        120, 122, 118, 121, 119, 123, 117, 124, 120, 122,   # Control group
        115, 118, 116, 117, 119, 114, 116, 118, 117, 115    # Treatment group
    ]
}

df = pd.DataFrame(data)

# Side-by-side boxplot with statistical annotation
ax = btp.boxplot_side_by_side(df, group_col="group", value_col="blood_pressure")
plt.show()

# Statistical test
x = df[df["group"] == "Control"]["blood_pressure"]
y = df[df["group"] == "Treatment"]["blood_pressure"]
result = btp.compare_two_groups(x, y)
print("Statistical Test Results:", result)

# Custom pastel colors (optional)
colors = ["#f0106dff", "#adf010ff"]  # pastel red and blue

ax = btp.boxplot_side_by_side(
    df,
    group_col="group",
    value_col="blood_pressure",
    colors=colors
)
plt.show()

Functions

boxplot_side_by_side(df, group_col, value_col)

• Creates a side-by-side boxplot for a two-level categorical variable.
• Automatically shows a pastel-colored plot with optional statistical annotation.

compare_two_groups(x, y)

• Performs Shapiro–Wilk normality tests on both groups.
• Runs t-test if both groups are normal, or Mann–Whitney U test if not.
• Returns a dictionary with test names and p-values.

License

MIT License. See LICENSE for details.

Contributing

Contributions, bug reports, and feature requests are welcome. Open an issue or pull request on GitHub.

Contributors

• Arkaprabha Sau (Author, Creator, Maintainer)
• Santanu Phadikar (Contributor)
• Ishita Bhakta (Contributor)