Charting a Whole Constellation: Fact Constellations and Galaxy Schemas

Opening Scene

A single star chart, however well designed, can’t hold everything a growing observatory tracks. Sales of telescope time, equipment maintenance events, and public viewing night attendance are each genuinely different processes, each deserving its own central star. But they’re not unrelated: they all still happen on the same shared calendar, involve the same shared equipment catalog, and take place at the same shared locations. The observatory’s full archive naturally becomes a constellation of several stars, connected not directly to each other, but through the shared points of light they all reference in common.

A galaxy schema, or fact constellation, is exactly this expanded structure.

In Plain English

A galaxy schema (also called a fact constellation) is what you get when multiple fact tables, each representing a genuinely different business process, share some of the same conformed dimensions. Rather than one isolated star schema per business process, a mature dimensional model typically ends up looking like several stars sharing common points of light — a natural, healthy outcome of building fact tables around genuinely conformed dimensions from the start.

The Old Way

Building toward a coherent galaxy schema has always been the natural, mature end-state of dimensional modeling done well over time, rather than something designed all at once:

Each fact table still represents one clear business process and one clear grain, exactly as covered earlier in this series — a galaxy schema doesn’t change this discipline, it just recognizes that most organizations genuinely have more than one business process worth modeling.
Shared, conformed dimensions are what make the constellation coherent, rather than just a loose pile of unrelated star schemas — the date, customer, and product dimensions covered in Article 8 are exactly what lets a galaxy schema support genuine drill-across analysis between different fact tables.
A galaxy schema typically emerges gradually, as an organization models one business process at a time and consciously reuses existing conformed dimensions rather than building a new, incompatible version each time — an intentional, disciplined outcome rather than an accident.

Getting this right has always meant treating dimension conformity as the genuine organizing principle across an entire dimensional model, not just within any single star schema considered in isolation.

What’s Changing (and Why AI Is the Reason)

AI-assisted schema mapping can visualize and analyze an organization’s full galaxy schema, surfacing which dimensions are genuinely shared and where conformity has quietly broken down across fact tables that should be connected. As a dimensional model grows to include many fact tables, understanding its full shape by hand becomes genuinely difficult; AI-assisted analysis can map the actual relationships and flag where dimensions that should be conformed have drifted apart.
AI agents answering cross-process business questions depend on a genuinely coherent galaxy schema to combine data from multiple fact tables correctly, making this pattern more consequential as agent-driven analysis grows. A question like “how does equipment maintenance downtime correlate with telescope time sales” can only be answered correctly if both fact tables genuinely share a conformed equipment or date dimension the agent can reliably join through.
AI-assisted new-fact-table design can actively recommend reusing existing conformed dimensions when a new business process is being modeled, rather than a team inadvertently building a redundant, incompatible version from scratch. This directly reduces the organizational drift that erodes galaxy schema coherence over time, catching the problem at its source rather than after the fact.

The Metaphor, Fully Extended

Observatory Element	Galaxy Schema Concept
Telescope time sales, equipment maintenance, and viewing night attendance, each its own star	Multiple fact tables, each representing a genuinely different business process
The shared calendar, equipment catalog, and location list every star chart references	Shared, conformed dimensions connecting multiple fact tables into one coherent galaxy
The observatory’s full archive naturally growing into several connected star charts over time	A galaxy schema emerging gradually as an organization models one business process at a time
An archive director mapping the full observatory archive to see how its charts genuinely connect	AI-assisted schema mapping visualizing an organization’s full galaxy schema and flagging drift
A new astronomer being pointed to the observatory’s existing shared calendar rather than starting a new one	AI-assisted new-fact-table design recommending reuse of existing conformed dimensions

For Beginners: What to Actually Do

Practice recognizing that most real, mature dimensional models are galaxy schemas — several connected star schemas — rather than a single isolated star.
Get comfortable with the idea that a galaxy schema’s coherence depends entirely on genuine dimension conformity, the concept covered in Article 8.
Before designing a new fact table, check whether an existing conformed dimension can be reused, rather than assuming a fresh version needs to be built.
Notice that a galaxy schema isn’t a different kind of modeling technique — it’s simply what a well-conformed dimensional model naturally looks like once it covers more than one business process.

For Practitioners and Leaders: The Deeper Layer

Use AI-assisted schema mapping to understand your organization’s full galaxy schema shape, surfacing where dimension conformity has genuinely broken down across fact tables that should be connected.
Design for AI agents’ growing role in cross-process business analysis explicitly, since they depend entirely on a coherent, genuinely conformed galaxy schema to answer multi-fact-table questions correctly.
Use AI-assisted new-fact-table design recommendations to catch redundant, incompatible dimension creation at the source, before it erodes galaxy schema coherence over time.
Treat dimension conformity as your organization’s genuine long-term modeling discipline, since it’s what determines whether your dimensional model matures into a coherent galaxy or a fragmented pile of incompatible stars.

Quick Recap

A galaxy schema, or fact constellation, is what results when multiple fact tables, each representing a different business process, share genuinely conformed dimensions.
This structure typically emerges gradually and intentionally, as an organization models one process at a time while consciously reusing existing conformed dimensions.
AI-assisted schema mapping can surface an organization’s full galaxy schema shape and flag where conformity has broken down, and AI-assisted design tooling can recommend reusing existing dimensions.
AI agents answering cross-process questions depend entirely on a genuinely coherent, conformed galaxy schema to combine data from multiple fact tables correctly.

Where This Fits in the Series

Article 14 covered whether a measurement actually adds up. This article covered charting a whole constellation. Article 16 looks at the number that doesn’t need its own planet — degenerate dimensions.