What data quality checks are performed?

The skill checks for data shape, column types, summary statistics, missing value patterns, duplicate rows, and distribution anomalies like outliers.

Does ds-plan require a project specification?

Yes, this skill follows the 'Iron Law of DS Planning' which requires a .claude/SPEC.md file to exist before profiling begins, ensuring all planning aligns with user objectives.

What is the output of this skill?

The primary output is a detailed .claude/PLAN.md file containing a data quality report and a step-by-step task breakdown for the implementation phase.

How does it handle large or multiple datasets?

It uses parallel background tasks to profile multiple data sources simultaneously, allowing for faster processing and consolidated reporting within the main chat.

DS Planning & Profiling

Name: DS Planning & Profiling
Author: edwinhu

byedwinhu

Data Science & ML

Profiles datasets and generates structured analysis task breakdowns for data science workflows.

About

The ds-plan skill is a specialized utility designed for the second phase of data science projects, focusing on rigorous data profiling and tactical planning. It prevents 'analysis drift' by forcing a comprehensive check of data shape, types, missing values, and quality issues before any modeling begins. By automating the creation of a structured .claude/PLAN.md file and supporting parallel profiling of multiple data sources, it ensures that implementation is based on reality rather than assumptions, significantly reducing rework and logic errors.

Key Features

Dependency-aware analysis task breakdown generation
Parallel execution of profiling tasks using background agents
0 GitHub stars
Comprehensive data profiling for single or multiple sources
Automated identification of missing values, duplicates, and outliers
Standardized .claude/PLAN.md documentation for project continuity

Use Cases

Bridging the gap between project requirements and technical implementation
Automating initial quality audits across disparate CSV or database sources
Establishing a data-validated roadmap for complex exploratory data analysis

About

Key Features

Dependency-aware analysis task breakdown generation
Parallel execution of profiling tasks using background agents
0 GitHub stars
Comprehensive data profiling for single or multiple sources
Automated identification of missing values, duplicates, and outliers
Standardized .claude/PLAN.md documentation for project continuity

Use Cases

Bridging the gap between project requirements and technical implementation
Automating initial quality audits across disparate CSV or database sources
Establishing a data-validated roadmap for complex exploratory data analysis