What is the difference between this and hypothesis testing?

Hypothesis testing checks a specific claim; Guided Investigation helps you discover what the claims should be by exploring data patterns systematically.

Does this skill require pre-processed data?

Yes, it requires data to be imported into a SQLite database and cleaned using the datapeeker importing and cleaning skills before starting.

When should I use the Guided Investigation skill?

Use it for open-ended questions where you lack a specific hypothesis, such as 'What is causing this trend?' or 'Why are we seeing these results?'

How does it help prevent unfocused analysis?

It enforces a decomposition phase where broad questions must be broken into 3-5 specific, answerable sub-questions before any querying begins.

Guided Investigation Framework

Name: Guided Investigation Framework
Author: tilmon-engineering

bytilmon-engineering

•

数据科学与机器学习

Conducts systematic, multi-phase investigations into complex, open-ended data questions using structured decomposition and incremental synthesis.

关于

The Guided Investigation skill provides a rigorous 5-phase methodology for tackling exploratory 'Why?' or 'What's driving X?' questions within Claude Code. It moves beyond simple querying by enforcing a workflow that includes question decomposition, data mapping, per-sub-question investigation, cross-finding synthesis, and actionable recommendations. By requiring explicit checkpoints and documentation at every stage, it ensures that data-driven investigations are focused, thorough, and auditable, preventing the common pitfall of chasing random patterns without a clear framework.

主要功能

Integrated data profiling and SQL query construction for SQLite databases
Incremental documentation and per-sub-question analysis tracking
2 GitHub stars
Systematic 5-phase investigation workflow from decomposition to recommendations
Structured sub-question mapping to prevent unfocused data exploration
Mandatory progress tracking and checkpoint verification using TodoWrite

使用场景

Exploratory data analysis of complex datasets with multiple variables
Root cause analysis for business metrics like customer churn or revenue drops
Generating structured strategic recommendations from raw data findings

关于

主要功能

Integrated data profiling and SQL query construction for SQLite databases
Incremental documentation and per-sub-question analysis tracking
2 GitHub stars
Systematic 5-phase investigation workflow from decomposition to recommendations
Structured sub-question mapping to prevent unfocused data exploration
Mandatory progress tracking and checkpoint verification using TodoWrite

使用场景

Exploratory data analysis of complex datasets with multiple variables
Root cause analysis for business metrics like customer churn or revenue drops
Generating structured strategic recommendations from raw data findings