{"categories":[{"label":"Information Analysis","url":"https://skillfed.io/packages/category/scientific-engineering-information-analysis/3"},{"label":"Mathematics","url":"https://skillfed.io/packages/category/scientific-engineering-mathematics/3"}],"enrichment":{"capability":"Provides post hoc statistical tests for pairwise multiple comparisons after ANOVA, including parametric tests (Scheffe, Student T, Tamhane T2, TukeyHSD) and non-parametric tests (Dunn, Nemenyi, Mann-Whitney, Wilcoxon) with p-value adjustment and outlier detection.","skillfed_tags":["statistics","hypothesis-testing","anova"],"use_cases":["Run post hoc pairwise comparisons after ANOVA to identify which group means differ significantly.","Apply non-parametric post hoc tests (Dunn, Nemenyi) when data violates normality assumptions.","Detect outliers in datasets using Grubbs, Tietjan-Moore, or ESD tests before analysis.","Generate significance plots to visualize pairwise comparison results for reports or publications.","Adjust p-values across multiple comparisons to control Type I error rates in exploratory studies.","Analyze block design experiments using specialized tests like Durbin-Conover or Quade."],"what_it_does":"scikit-posthocs fills a gap in Python's statistical ecosystem by offering a comprehensive suite of post hoc tests for pairwise multiple comparisons\u2014the follow-up analysis performed after ANOVA or similar omnibus tests reveal statistical significance. It implements both parametric tests (like TukeyHSD and Scheffe) and non-parametric alternatives (like Dunn and Nemenyi tests) for different experimental designs, plus outlier detection methods and basic plotting. The package is tightly integrated with pandas DataFrames and NumPy arrays, making it convenient to work with typical data science workflows.\n\nYou would use this package when you've run an ANOVA or Kruskal-Wallis test and need to determine which specific group pairs differ significantly. It handles p-value adjustment automatically to control for multiple comparisons, and it supports both factorial and block designs. The package depends on scipy, statsmodels, pandas, seaborn, and matplotlib, so it fits naturally into existing scientific Python environments.","worth_installing":"Yes. The package is actively maintained, has no known vulnerabilities, uses a permissive MIT license, and fills a genuine gap in Python's statistical toolkit by providing post hoc tests that are either missing or inconvenient in scipy and statsmodels. Install friction is low, and it integrates well with standard data science libraries. Recommended for researchers, statisticians, and data analysts working with group comparisons."},"id":"scikit-posthocs","links":{"html":"https://skillfed.io/packages/scikit-posthocs","md":"https://skillfed.io/packages/scikit-posthocs.md","pypi":"https://pypi.org/project/scikit-posthocs/"},"maintenance":{"status":"active"},"meta":{"latest_release":"2026-05-26","license_spdx":null,"license_treatment":"permissive","name":"scikit-posthocs","python_support":"supports_current","summary":"Statistical post-hoc analysis and outlier detection algorithms"},"popularity":{"monthly_downloads":146623,"position":11093,"tier":"top_15000"},"security":{"n_vulnerabilities":0},"version":"0.14.0"}