semantic-model-builder

nimrodfisher's avatarfrom nimrodfisher

Create comprehensive semantic layer documentation for analytics assets. Use when documenting data models, defining business metrics, creating data dictionaries, or building context for AI-assisted analysis.

⭐0stars🔀0forks📁View on GitHub🕐Updated Jan 11, 2026
[SQL and Databases]

When & Why to Use This Skill

The Semantic Model Builder is a specialized Claude skill designed to bridge the gap between raw data and business intelligence. It automates the creation of comprehensive semantic layer documentation, enabling teams to define business metrics, document data models, and build structured data dictionaries. By providing clear context and calculation logic, it ensures a single source of truth for analytics and significantly enhances the accuracy of AI-assisted data analysis.

Use Cases

  • Metric Standardization: Define complex KPIs like MRR, DAU, or Conversion Rates with precise SQL logic and business context to ensure consistency across the organization.
  • Data Dictionary Creation: Automatically generate detailed documentation for database tables, including column descriptions, primary keys, and entity-relationship mappings.
  • AI Context Building: Prepare structured metadata and semantic layers that allow LLMs to understand and query underlying data structures with higher precision.
  • Data Governance & Onboarding: Create a centralized knowledge base of business concepts and data models to facilitate faster onboarding for new analysts and maintain data integrity.
namesemantic-model-builder
descriptionCreate comprehensive semantic layer documentation for analytics assets. Use when documenting data models, defining business metrics, creating data dictionaries, or building context for AI-assisted analysis.

Semantic Model Builder

Quick Start

Build structured documentation that defines business metrics, data models, and relationships in a format optimized for AI-assisted analysis.

Context Requirements

  1. Metric/Entity to Document: What needs documentation
  2. Calculation Logic: How it's computed (SQL, formula, or plain English)
  3. Business Context: Why it matters, how it's used
  4. Data Sources: Where the data comes from

Context Gathering

Initial Prompt:

"Let's build semantic documentation. What would you like to document?

  • A specific metric (e.g., MRR, DAU, Conversion Rate)
  • A data model/table (e.g., users table, transactions)
  • A business concept (e.g., 'Active Customer')
  • Multiple related items"

For Metrics:

"For [metric name], I need:

  1. Definition: What is this metric in plain English? Example: 'Monthly Recurring Revenue (MRR) is the predictable revenue generated each month from active subscriptions'

  2. Calculation: How is it calculated?

    • Provide SQL query, OR
    • Formula (e.g., 'SUM(subscription_amount) WHERE status = active'), OR
    • Plain English steps

  3. Business Context:

    • Why does this metric matter?
    • Who uses it?
    • What decisions does it inform?
    • What's a 'good' value?

  4. Edge Cases (optional but helpful):

    • What should be included/excluded?
    • How to handle special situations?
    • Known calculation gotchas?"

For Data Models:

"For [table/model name], I need:

  1. Purpose: What does this table represent? Example: 'One row per user signup'

  2. Key Columns: Most important fields

    • Which are IDs/keys?
    • Which are metrics?
    • Which are attributes?

  3. Relationships: How does this connect to other tables? Example: 'users.id → orders.user_id'

  4. Grain: What is one row? Example: 'One row per transaction' or 'One row per user per day'"

For Business Concepts:

"For [concept], help me understand:

  1. Definition: What is this?
  2. How to Identify: How do you know something is/isn't this?
  3. Related Data: Where is this captured in data?
  4. Why It Matters: Business significance?"

Workflow

1. Gather Information

Start with what's provided, probe for gaps:

If user provides SQL: