搜索

PythonReactSEOPDF
GitHub Skill

ydata-eda-profiling

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.

crossxwill☆ 5 仓库 Stars0 仓库 Forks

安装与使用

请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求,再添加到工作流。

查看来源 ↗使用指南

技能说明

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.

使用示例

查看来源仓库中的使用示例与配置说明。

查看来源 ↗

来源历史

收录时间
2026-01-20
索引更新时间
2026-01-26
Branch
main

收录与同步日期属于本站记录。源码修改历史请查看来源仓库。

GitHub ↗

chat Comments (0)

chat_bubble_outline

No comments yet. Be the first to share your thoughts!