Design data science agent systems

What problem does it solve?

This Skill addresses the challenge of creating data science agent systems that avoid fabricating data, manage uncertainty effectively, and provide actionable insights, even in complex analytical scenarios.

Core Features & Use Cases

• Anti-Fabrication Rules: Ensures agents never create false data, crucial for accurate analysis.
• Retry Protocols: Guides agents to retry and simplify analyses when models fail, maintaining methodological rigor.
• Conflict Detection: Helps identify conflicting results across diverse approaches, indicating underlying issues.
• Epistemic Humility: Instructs agents to recognize when to stop analysis and report uncertainty transparently.
• Prompt Design Principles: Encourages clear, non-implementational prompts to enable reasoning and learning.
• Use Case: Use this Skill to develop a robust marketing mix modeling agent that can handle multiple approaches and provide insightful, convergent recommendations.

Quick Start

Create an agent system using the Design data science agent systems skill to perform a multi-model analysis on the marketing dataset 'mmm_data.csv'.