Does it provide guidance on sample size?

Yes, the skill includes power analysis patterns to help determine appropriate sample sizes or interpret negative results effectively.

What Python libraries does this skill use?

It leverages industry-standard Python libraries including Pandas, NumPy, SciPy, Matplotlib, Seaborn, and Scikit-learn for comprehensive analysis.

Does this skill handle non-normal data distributions?

Absolutely. It provides a decision tree and code for non-parametric tests like Mann-Whitney U and Kruskal-Wallis when normality assumptions are violated.

Can it help with multiple testing problems?

Yes, the skill includes specific implementations for Bonferroni and False Discovery Rate (FDR) corrections using statsmodels.

Is it suitable for high-dimensional data visualization?

Yes, it features standardized workflows for Principal Component Analysis (PCA) and t-SNE to help visualize structure in high-dimensional datasets.

Data Science for Scientific Discovery

Name: Data Science for Scientific Discovery
Author: justaddcoffee

byjustaddcoffee

データサイエンスとML

Provides robust statistical analysis strategies and data exploration techniques for rigorous scientific research and discovery.

概要

This skill equips Claude Code with a comprehensive framework for scientific data analysis, guiding users through initial data assessment, distribution validation, and the selection of appropriate statistical tests. It offers implementation patterns for complex tasks like multiple testing correction, dimensionality reduction, and confounder adjustment, ensuring that data-driven insights are statistically sound and scientifically reproducible. Whether you are running t-tests or performing unsupervised clustering, this skill ensures best practices in data science are applied to your scientific workflows.

主な機能

Automated initial data assessment and distribution checking
Multiple testing correction (FDR, Bonferroni) implementations
Dimensionality reduction techniques including PCA and t-SNE
Power analysis and confounder adjustment strategies
Statistical test decision tree for hypothesis testing
0 GitHub stars

ユースケース

Interpreting complex scientific datasets using clustering and PCA
Validating statistical assumptions before running hypothesis tests
Correcting for multiple comparisons in high-throughput screening data

概要

主な機能

Automated initial data assessment and distribution checking
Multiple testing correction (FDR, Bonferroni) implementations
Dimensionality reduction techniques including PCA and t-SNE
Power analysis and confounder adjustment strategies
Statistical test decision tree for hypothesis testing
0 GitHub stars

ユースケース

Interpreting complex scientific datasets using clustering and PCA
Validating statistical assumptions before running hypothesis tests
Correcting for multiple comparisons in high-throughput screening data