Back to AnalyzerHow It Works

Understanding Diversity Analysis

Learn how we measure variety and uniqueness in AI-generated floor plans to ensure you're exploring the full design possibility space.

The Simple Version

Imagine you asked 10 architects to design a 3-bedroom house. Would you want 10 nearly identical designs, or 10 genuinely different approaches?

AI can sometimes get "stuck" generating similar designs. This tool measures how different your generated floor plans actually are from each other.

Low Diversity (Bad)

All plans look similar - same room arrangements, same flow patterns. The AI isn't exploring creative alternatives.

High Diversity (Good)

Each plan takes a different approach - varied layouts, room positions, and circulation patterns. More options to choose from!

We give you a diversity score from 0-100%. Higher means more variety. We also show you a scatter plot where each dot is a floor plan - clustered dots mean similar designs, spread out dots mean diverse designs.

What We Analyze

Room Distribution

How rooms are sized and proportioned. Are bedrooms similar sizes, or is there variety?

Circulation Patterns

How you move through the space. Linear hallways vs. open flow vs. central hubs.

Spatial Relationships

Which rooms connect to which. Kitchen next to dining? Bedrooms grouped or separated?

Overall Massing

The shape and footprint. Compact squares, L-shapes, sprawling layouts, or something unique.

Reading the Scatter Plot

The scatter plot is a "map" of your designs. Each dot represents one floor plan, positioned based on its features.

  • Dots close together = Similar designs
  • Dots spread apart = Diverse designs
  • Colored groups (clusters) = Plans that share similar characteristics

If all your dots are in one tight cluster, the AI is stuck in one "design mode." If dots are spread across the space, you're getting genuine variety.