ydata-eda-profiling

name: ydata-eda-profiling description: Generate and compare ydata-profiling EDA reports with sampling, consistent random seeds, and HTML outputs; often follows duckdb-parquet-lab-workflow when data is queried from Parquet.

YData Profiling EDA

Purpose

Create consistent EDA reports for train and test or accepted and rejected datasets using ProfileReport, including comparisons and saved HTML outputs.

Usage

• "generate ydata-profiling report"
• "compare train and test EDA"
• "create HTML EDA report"

Instructions

1. Set a sampling fraction for large datasets and a fixed random seed.
2. Create ProfileReport objects with progress_bar=False, and disable duplicates and interactions when speed matters.
3. Compare reports using .compare() and save to HTML with .to_file().
4. Use ./scripts/generate_eda_report.py for consistent report creation.
5. Use ./templates/eda_compare_block.md to document inputs, sample fractions, and output paths.