data-scientist

What it does

Guides a data-science workflow for Python research, covering EDA, data validation, transformation verification, documentation standards, visualization design, descriptive analysis, statistical modeling, causal inference method selection (IV, DiD, RD, synthetic control), unsupervised analysis (clustering, PCA, UMAP), supervised ML methodology, and geospatial analysis. Provides methodology decisions and analytical approach guidance. For implementation syntax, load the tool-specific skill (polars, statsmodels, plotnine, pyfixest, scikit-learn, geopandas, etc.). Use for any data analysis, exploration, transformation, or modeling task — especially when choosing methods, checking assumptions, or structuring an analysis.

Establishes a rigorous, methodical approach to data science work. This skill is about how to think and work, not specific tools. Load specialized skills (polars, plotnine, plotly, marimo, etc.) for tool-specific guidance.

How it works

• Emphasizes Core Principles that govern all work: Data Robustness First (check data before operating), Documentation First (understand or create data documentation), Verify Every Operation (check counts, samples, and outcomes), Thorough Code Documentation (Inline Audit Trail standards), and Focus on Research Questions (balance rigor with usefulness).
• Recommends loading core workflow connectors: polars for DataFrame operations; marimo for creating validated notebooks; and various domain-specific skills for visualization, modeling, and geospatial tasks.
• Provides a multi-step guidance structure for loading and using domain-specific reference materials before applying tool-specific syntax skills.

When to use it

Use when starting any data analysis project to establish methodology, when selecting causal, ML, or unsupervised approaches, and when structuring a research workflow that benefits from rigorous data handling, documentation, and validation. It instructs loading domain-specific reference materials and tool-skills in a prescribed order to ensure principled implementation.

What it can touch

• Prerequisite knowledge and tool loading relationships are described, including loading polars, marimo, plotnine, plotly, scikit-learn, geopandas, statsmodels, pyfixest, and others as needed for syntax and implementation.
• The skill references a collection of reference files (e.g., eda-checklist.md, data-documentation.md, etc.) to guide analysis steps, validations, and reporting.

Caveats

• License listed as NOASSERTION.
• This skill provides methodology guidance; specialized skills provide the actual implementation syntax.