Senior Data Scientist — Data

This skill provides a comprehensive toolkit for senior data science tasks, including A/B testing, feature engineering, model evaluation, and causal inference. It includes Python code templates, checklists, and best practices to guide users through designing experiments, building pipelines, and interpreting results.

engineering claude-code

📂 View skill on GitHub ↗ 🚀 How to use

✨ What it does

Provides sample size calculation and two-proportion z-test analysis for A/B tests.
Includes checklists for experiment design, feature engineering, model evaluation, and causal inference.
Offers Python code for building feature pipelines with scikit-learn and adding time-based features.
Supports cross-validated model evaluation with AUC-ROC, AUC-PR, and SHAP values.
Integrates MLflow for experiment tracking and logging.
Implements difference-in-differences estimation with robust standard errors.

🎯 When to use it

When designing or analyzing controlled experiments such as A/B tests.
When building and evaluating classification or regression models on structured data.
When performing causal analysis on observational data using methods like difference-in-differences.
When engineering features for tabular datasets to improve model performance.
When translating statistical findings into data-driven business decisions.

🚀 How to use

To use this skill, provide a request that specifies the data science task you need help with, such as designing an A/B test, building a feature pipeline, evaluating a model, or performing causal analysis. The skill will generate code snippets and checklists tailored to your input. For example, you can ask: 'Help me design an A/B test for a new checkout flow with a baseline conversion rate of 10% and a desired lift of 5%.' or 'I need to evaluate an XGBoost model on an imbalanced dataset; what metrics should I use?' The skill expects you to provide relevant parameters like baseline rates, effect sizes, column names, or dataset details.

📄 Output: The skill produces code snippets, checklists, and guidance for the requested data science task, which can be directly used in your project.

📦 Add this skill to Claude Code

# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills

# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/engineering-team/skills/senior-data-scientist .claude/skills/senior-data-scientist

Skill source: engineering-team/skills/senior-data-scientist/SKILL.md

⚠️ Good to know

The skill assumes familiarity with Python, statistics, and machine learning concepts; it provides templates and best practices but does not execute code or handle data directly.

❓ FAQ

What is the recommended sample size calculation method?

The skill provides a function to calculate sample size per variant based on baseline rate, minimum detectable effect, alpha, and power, using a normal approximation.

How should I handle multiple metrics in an A/B test?

Apply Bonferroni correction by dividing the alpha level by the number of metrics tested to control the family-wise error rate.

What metrics should I report for imbalanced classification problems?

Always report AUC-PR alongside AUC-ROC, as AUC-PR is more informative for imbalanced datasets.

🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.