Three Modeling Methodologies

Inmon, Kimball, and Data Vault 2.0

Bill Inmon

The "Father of Data Warehousing"
Top-Down · 3NF
  • Enterprise-wide normalized warehouse
  • Single source of truth (3NF)
  • Data marts derived from the warehouse
  • Long initial build, high consistency
  • Best for: large enterprises needing one governed model

Ralph Kimball

Dimensional Modeling Pioneer
Bottom-Up · Star Schema
  • Business-process-oriented data marts
  • Star schema: facts + dimensions
  • Conformed dimensions for integration
  • Fast delivery, BI-optimized
  • Best for: BI-first teams wanting quick value

Data Vault 2.0

Dan Linstedt
Hub-Link-Satellite · Agile
  • Hubs (business keys), Links (relationships), Satellites (descriptive context)
  • History-preserving, fully auditable
  • Parallel loading from many sources
  • Agile and scalable for integration
  • Best for: complex multi-source environments

Bill Inmon: Top-down enterprise warehouse

WhatWhyHow
What it is

Inmon advocates building a centralized, normalized (3NF) enterprise data warehouse first, then deriving dimensional data marts for specific business units. The warehouse is the single source of truth, and all downstream models inherit its consistency.

Why choose this

When the organization needs a single, consistent, enterprise-wide view of data. 3NF minimizes redundancy and update anomalies. Ideal for heavily regulated industries (banking, healthcare) where a single governed truth is critical.

Trade-offs
  • Pro: Maximum consistency, minimal redundancy, strong governance
  • Pro: Well-suited for complex, evolving business rules
  • Con: Long initial build time (can take months before first value)
  • Con: 3NF queries require many joins, which can be slow for BI
  • Con: Requires strong upfront enterprise data modeling expertise

Ralph Kimball: Bottom-up star schema

WhatWhyHow
What it is

Kimball advocates building dimensional data marts (star schemas) organized by business process. Facts tables hold measures (revenue, quantity, clicks). Dimension tables hold descriptive context (customer, product, date). Conformed dimensions link marts together into an integrated warehouse.

Why choose this

Fastest path to BI value. Star schemas are intuitive, query-friendly, and optimized for the way analysts think. Build one mart at a time, deliver incremental value, and connect marts through shared dimensions.

Key concepts
  • Fact table: stores measurable events (sales, clicks, shipments). Contains foreign keys to dimensions and numeric measures.
  • Dimension table: descriptive context (who, what, where, when). Customer, Product, Date, Geography.
  • Star schema: one fact table surrounded by dimensions. Simple, fast joins.
  • Snowflake schema: normalized dimensions (dimension references sub-dimensions). Saves space but adds joins.
  • Conformed dimensions: shared dimensions across multiple fact tables for consistency.

Data Vault 2.0: Hubs, links, and satellites

WhatWhyHow
What it is

Data Vault 2.0 (Dan Linstedt) is a modeling methodology designed for agile, scalable data integration from many sources. It separates data into three entity types: Hubs (business keys), Links (relationships between hubs), and Satellites (descriptive and historical context).

Why choose this

When you need to integrate data from many disparate sources, preserve full history, and support agile development. Data Vault is uniquely suited for environments where sources change frequently and full auditability is required.

Components
  • Hub: contains only business keys (e.g., customer_id, product_sku). Immutable once loaded.
  • Link: represents relationships between hubs (e.g., customer_order link). Also immutable.
  • Satellite: holds descriptive attributes and their history (e.g., customer address history). Temporal, with load_date and record_source.
  • Pattern: Sources load in parallel to hubs and satellites. No need to coordinate cross-source loading.
DimensionInmon (3NF)Kimball (Star)Data Vault 2.0
ApproachTop-downBottom-upHub-link-satellite
StructureNormalized (3NF)Denormalized (star/snowflake)Hubs + Links + Satellites
Primary goalSingle source of truthBI-optimized fast deliveryScalable, auditable integration
Time to first valueLong (months)Short (weeks)Medium
History trackingRequires extra designSCD typesBuilt-in (satellites)
ComplexityHigh (modeling)Low to moderateHigh (many tables)
Best forLarge regulated enterprisesBI-first analytics teamsMulti-source integration at scale

Modern trend: hybrid approaches

What teams actually do

In practice, many modern data teams combine approaches: Data Vault for the integration layer (Silver, where multiple sources merge with full history) and Kimball star schemas for the consumption layer (Gold, optimized for BI and analytics). This gives you the scalability and auditability of Data Vault where sources land, plus the query performance and analyst friendliness of star schemas where dashboards connect.

ETL vs. ELT and Schema Strategies

Where and when to apply structure

Interactive: ETL vs. ELT decision

Select a scenario to see which approach fits.

ETL: Extract, Transform, Load

How it works: Data is extracted from sources, transformed in a separate processing engine (Spark, Python, Informatica), then loaded into the target warehouse. Best for: Environments where the warehouse has limited compute, or when sensitive data must be masked/filtered before loading. Trade-off: Transformation logic lives outside the warehouse, making it harder to iterate and test with SQL.

Slowly Changing Dimensions (SCD)

How to handle changes in dimensional data

SCD types: preserving or overwriting history

TypeStrategyHistory?Example
SCD Type 0Never changeN/A (immutable)Date of birth, original signup date
SCD Type 1OverwriteNo (lost)Fix a typo in a customer name
SCD Type 2Add new row with validity datesYes (full history)Customer changes address: old row gets end_date, new row gets start_date
SCD Type 3Add column for previous valuePartial (one change)current_city and previous_city columns
Key insight

SCD Type 2 is the most important for data engineers. It preserves full history by creating a new row for each change, with effective_from and effective_to dates plus a current_flag. This lets analysts answer "what was the customer's address when they placed this order?" rather than only "what is the customer's address now?"

Knowledge Check: Module 5

1. Which modeling methodology builds dimensional star schemas organized by business process?

Kimball's approach builds dimensional data marts (star schemas) organized by business process, with fact tables holding measures and dimension tables holding descriptive context. Conformed dimensions link marts together.

2. What are the three entity types in Data Vault 2.0?

Data Vault 2.0 uses Hubs (business keys), Links (relationships), and Satellites (descriptive attributes with history). This separation enables parallel loading and full auditability.

3. What is a fact table in dimensional modeling?

Fact tables hold numeric measures (revenue, quantity, clicks) and foreign keys to dimension tables. Each row represents a business event (a sale, a click, a shipment).

4. Which SCD type preserves full history by adding new rows with validity dates?

SCD Type 2 creates a new row for each change, with effective_from and effective_to dates. This preserves the complete history of changes so you can answer point-in-time questions.

5. What is the modern hybrid modeling approach many teams use?

The hybrid approach uses Data Vault where sources merge (scalable, auditable integration) and Kimball star schemas where consumers query (fast, intuitive BI). This combines the strengths of both methodologies.