Quantitative Analysis of Contemporary Residential Bathroom Design Through AI Visual Recognition: Object-Design Characteristic Mapping of 1,500 Cases

Abstract

This study presents a novel methodology for quantitatively analyzing contemporary residential bathroom design characteristics by combining Large Language Model (LLM) image recognition capabilities with text mining techniques. We collected 1,500 bathroom images from ArchDaily and analyzed 1,492 valid cases using Claude API, applying natural language processing and topic modeling to the generated text data. Our analysis reveals a remarkable near 1∶ 1 balance between material nouns (physical objects) and abstract nouns (design attributes), empirically demonstrating that contemporary bathrooms have evolved from purely functional spaces to venues for aesthetic self-expression. We identified “sophisticated modern minimalism” as the dominant design language, characterized by achromatic palettes and refined simplicity. Through Latent Dirichlet Allocation (LDA) topic modeling, we uncovered six major design themes: Luxury Modern, Nature-Friendly, Functional Vanity Space, Lighting/Open-concept, Industrial, and Minimalism. This research demonstrates the viability of AI-based architectural image analysis and presents methodological innovations by applying discovery science approaches to architectural design research.