statistical-analysis

Hypothesis testing requires verifying assumptions like normality and variance homogeneity first. This skill automates these diagnostic checks using scipy and pingouin, then guides you to the correct parametric or non-parametric test and post-hoc analysis path.

APA-formatted reporting translates raw statistical output into academic standard text. This skill generates APA-formatted results and publication-ready visualizations automatically using matplotlib and seaborn after completing your hypothesis tests or regression diagnostics.

Bayesian analysis workflows are fully supported alongside frequentist methods. The skill uses pymc and arviz for Bayesian modeling while leveraging statsmodels for comprehensive regression diagnostics, allowing you to compare both approaches for your research data.

Selecting the correct statistical test involves matching your research question and data structure to the appropriate method. This skill provides guided test selection, navigating from t-tests to advanced models by evaluating your dataframe columns and grouping variables.

Pandas and numpy are required dependencies for data manipulation and numerical computation. You also need scipy, statsmodels, pingouin, and pymc installed to execute the full range of parametric, non-parametric, and Bayesian analysis workflows.

Non-parametric tests are necessary when your data fails normality or variance assumptions. This skill automates assumption checking and seamlessly transitions to non-parametric alternatives if your diagnostic checks indicate parametric tests are inappropriate.