Examples

Three runnable Jupyter notebooks ship with the package under examples/ — each one executes end-to-end in CI and is the canonical reference for its workflow.

Notebook	Workflow
01_basic_ttest.ipynb	`run_test("ttest_ind", …)` → unified result dict → APA string
02_test_recommendation.ipynb	`StatContext` → `recommend_tests` → top recommendation through `run_test`
03_multiple_comparison.ipynb	Family of comparisons → `correct.correct_fdr` (Benjamini-Hochberg)

To re-execute every notebook in place (refreshing outputs):

bash examples/00_run_all.sh

Example 1 — Basic t-test

Example t-test visualization — **Figure 1.** Box plot with individual data points, significance bracket, p-value, and effect size — generated from the unified result dictionary.

import numpy as np
import scitex_stats as ss

rng = np.random.default_rng(42)
group1 = rng.normal(0.0, 1.0, 30)
group2 = rng.normal(0.5, 1.0, 30)

result = ss.run_test("ttest_ind", data=group1, data2=group2)

assert result["stat_symbol"] == "t"
assert result["effect_size_metric"] == "Cohen's d"
print(result["formatted"])
# → t = -3.2101, p = 0.0022, Cohen's d = -0.829, **

Example 2 — Automatic test recommendation

import scitex_stats as ss

ctx = ss.StatContext(
    n_groups=2,
    sample_sizes=[30, 30],
    outcome_type="continuous",
    design="between",
    paired=False,
)
ss.recommend_tests(ctx, top_k=3)
# → ['ttest_ind', 'welch_t', 'brunner_munzel']

Example 3 — Multiple comparison correction

from scitex_stats import correct

results = [
    {"pvalue": 0.010, "var_x": "A", "var_y": "B"},
    {"pvalue": 0.040, "var_x": "A", "var_y": "C"},
    {"pvalue": 0.030, "var_x": "A", "var_y": "D"},
    {"pvalue": 0.200, "var_x": "B", "var_y": "C"},
    {"pvalue": 0.005, "var_x": "B", "var_y": "D"},
    {"pvalue": 0.080, "var_x": "C", "var_y": "D"},
]

corrected = correct.correct_fdr(results, alpha=0.05, method="bh")
# → BH-adjusted p-values + rejection mask + α_adjusted per comparison