Contents

Semantic analysis of websites for architectural design feature extraction

Author(s): M. Zhao1, J. Liang2, P. Shen2, E. Lange3, X. Ye4
1The School of Architecture and Urban Planning, Guangzhou University
2China Urban Construction Design & Research Institute Co., Ltd.
3University of Bristol, UK
4Stevens Institute of Technology
M. Zhao
The School of Architecture and Urban Planning, Guangzhou University
J. Liang
China Urban Construction Design & Research Institute Co., Ltd.
P. Shen
China Urban Construction Design & Research Institute Co., Ltd.
E. Lange
University of Bristol, UK
X. Ye
Stevens Institute of Technology

Abstract

This study examines how perceived architectural design attributes and evaluation themes can be inferred from website-based discourse on building design. We analyse several thousand online comments and propose a replicable lexical-semantic pipeline for summarizing how designs are discussed in natural online settings. The approach combines (i) corpus construction from professional discussions on a Chinese architectural forum (ABBS), (ii) filtering against a modern Chinese reference corpus to suppress general-language terms, and (iii) term salience ranking to support comparison among alternative schemes. The method is demonstrated on public comments about competing design proposals for the Zhangjiakou Olympic Stadium. Results indicate that online discourse consistently foregrounds descriptors related to exterior and form, functional usability, environmental performance, and cost/practicality, and that the relative emphasis of these descriptors differs across proposals. The approach provides a scalable complement to conventional survey-based evaluation, while acknowledging limitations regarding phrase-level meaning, context, and the representativeness of online commenters.

Copyright © 2024 M. Zhao, J. Liang, P. Shen, E. Lange, X. Ye. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.