data-modeling

name	data-modeling
description	Designs and validates normalized or dimensional data models. Assesses schema design for analytics readiness, identifies grain, cardinality, and fact vs. dimension patterns. Outputs entity-relationship diagram and modeling documentation.
allowed-tools	Read, Write, WebFetch
effort	high

When to activate

When designing a new analytics data warehouse, evaluating existing schema for analytics use cases, or refactoring data models to support new metrics. Run before building dbt models or star schema implementations.

When NOT to use

Not for SQL optimization — use sql-optimizer skill instead. Not for business logic implementation without understanding the underlying data structure first. Not without clear requirements about analytical grain and fact tables.

Modeling Checklist

1. Define analytical grain — Identify the atomic level of measurement (transaction, event, daily user, etc.)
2. Identify fact tables — Tables that measure business events with foreign keys to dimensions
3. Identify dimension tables — Tables that describe entities (customers, products, dates, etc.)
4. Check cardinality — Understand primary key relationships and fan-out patterns
5. Verify normalization — Assess if 3NF, star schema, or denormalized design is appropriate
6. Flag snowflaking — Detect unnecessary dimension hierarchy levels that could be flattened