Data Center Generator

I like it when tools make the design clearer, not noisier.

This is a small slice of how I approach data centers: take the logic we use every day (program, constraints, redundancy, flows), make it explicit, then let automation do the repetitive part. It’s simple rules, modeled in a way that can evolve.

The idea

• Inputs: program targets, site limits, tier, IT load
• Strategies: cooling and power, sized to the program
• Layout logic: generate data halls, corridors, mechanical yard (chillers), electrical yard (generators), and support zones
• Output: geometry blocks you can place, revise, and re-run

Why do this?
Because it turns “tribal knowledge” into something we can test, share, and improve. If the program changes, you re-run it. If a rule is wrong, you fix the rule—once.

Facility-level view (high-level flow: program → strategies → layout → baked geometry)

What you’re seeing: the structure of the generator, not the secret sauce.

• Program + site → cooling & power strategies
• Strategies feed a layout generator
• Generator places data halls, corridors, mech yard (chillers), elec yard (generators), support
• Then it bakes the geometry

Zoomed-in demo (smaller example to show the pattern without all the parts)

Same pattern, fewer blocks. Enough to see how inputs move through decisions into geometry.

What this enables

• Faster iteration: change a number, re-run, see the impact
• Consistency: rules live in code, not in someone’s memory
• Focus: teams spend time on decisions, not clicking through the same steps
• Change: as needs shift, the logic updates and the outputs follow

This is the kind of work I like—practical automation that helps teams ship better designs with less guesswork.

Tech Stack:
Dynamo, Revit API, Python