statistical-analysis

This skill serves as an end-to-end companion for researchers and data analysts performing systematic hypothesis testing and data interpretation. It provides an automated workflow to determine appropriate statistical methodologies based on research questions, variable types, and distribution characteristics. Designed for academic and clinical environments, the agent helps users navigate from raw data exploration to professional, publication-ready reporting.

• Intelligent test selection guide for comparing groups (t-tests, ANOVA, Mann-Whitney U, Wilcoxon, Kruskal-Wallis) and relationships (correlation, linear/logistic regression).

• Comprehensive assumption checking suite including outlier detection, normality tests (Shapiro-Wilk), homogeneity of variance (Levene's test), and linearity diagnostics.

• Support for Bayesian statistical alternatives, providing Bayes Factors and probabilistic interpretations to supplement frequentist results.

• Automated generation of APA-formatted statistical reports, including tables, figure captions, and clear, narrative interpretations.

• Integrated power analysis capabilities for planning studies and determining required sample sizes based on effect sizes and alpha levels.

• Users should provide clean, structured dataframes (e.g., in pandas) along with clearly defined research objectives to receive precise test recommendations.

• The skill leverages specialized Python libraries, including scipy.stats for core tests, statsmodels for complex regression, pingouin for streamlined statistical reporting, and pymc/arviz for Bayesian modeling.

• Always perform diagnostic visualizations, such as Q-Q plots and residual plots, provided by the built-in assumption check scripts before finalizing any analysis.

• For programmatic model implementation or custom script construction, combine this skill with standard coding assistance, as this tool focuses on the selection, verification, and interpretation pipeline.

• Note that this skill is optimized for research-grade rigor; it prioritizes transparent assumption verification to ensure the validity of the resulting statistical conclusions.