Pingouin Excel Statistical Analysis

Trigger

• •Use for statistical analysis tasks (hypothesis tests, post-hoc tests, and summary tables) where the final deliverable is an Excel workbook.
• •Trigger for requests mentioning either pingouin or typo variants like pinguoin.
• •In analysis/, if users want to try statistical analysis or aggregation directly, treat notebook-first exploration plus Python-script Excel generation as the preferred pattern.
• •For step vs nonstep EMG rank interpretation workflows in analysis/, co-use analysis-step-nonstep-rank-report and let that skill drive folder/report structure while this skill owns tests and Excel table rules.
• •In this repository, for statistical ranking/hypothesis-result requests (even when users do not explicitly say "Excel"), default deliverables should include:
  • •reproducible .py analysis code, and
  • •an .xlsx workbook (tables + table_guide) for review/validation.

Non-Negotiable Rules

• •Use pingouin for statistical testing (friedman, anova, pairwise_tests, and related functions).
• •Prefer polars for data preparation, then convert to pandas only when a statistics API requires it.
• •Write Excel outputs with xlwings in Windows PowerShell or cmd.
• •Use .xlsx as the default output format (use .xlsm only when macro requirements are explicitly requested).
• •For analysis/ statistical or aggregation tasks, prefer:
  • •.ipynb for exploratory aggregation/statistical iteration.
  • •.py for final Excel workbook generation.
  • •This is a preferred workflow, not a global hard requirement for every task.
• •Save user-facing results as Excel Tables (ListObject), not loose ranges.
• •Keep sheet count minimal. Default layout:
  • •tables: all result tables.
  • •table_guide: metadata and explanation for each table.
• •Every output table must have a matching table_guide row with a non-empty description.

Table Guide Schema

• •table_name
• •sheet_name
• •table_range
• •description
• •key_columns
• •filters_or_notes

Workflow

1. •Define hypotheses and required output tables before coding.
2. •In analysis/ statistical or aggregation exploration, iterate in .ipynb first when useful.
3. •Load and clean data in polars.
4. •Convert only required frames to pandas.
5. •Compute statistics with pingouin.
6. •Move final Excel-generation logic into a .py script.
7. •Write each result block to tables and convert ranges to Excel Tables.
8. •Populate table_guide with interpretation notes.
9. •Reopen workbook and validate errors, blanks, and table-guide integrity.

Validation

• •Fail if any cell contains #DIV/0!, #N/A, #NAME?, #NULL!, #NUM!, #REF!, or #VALUE!.
• •Fail if expected tables are missing.
• •Fail if table_guide is missing or has blank description values.
• •Fail if business-critical required ranges are blank.

Command Templates

NOTE: scripts/run_stats_report.py is a placeholder example path (this repo may not include that file). Replace it with your actual stats-report script under scripts/.

PowerShell:

• •conda run -n module python -c "import pingouin, xlwings; print('ok')"
• •conda run -n module python .\scripts\run_stats_report.py

cmd:

• •conda run -n module python -c "import pingouin, xlwings; print('ok')"
• •conda run -n module python scripts\run_stats_report.py

Completion Checklist

• •Report workbook path.
• •Report final sheet count and rationale if more than two sheets.
• •List created table names.
• •Confirm table_guide descriptions were added for all tables.
• •Report validation results.