Social Media Data in Urban Design and Landscape Research: A Comprehensive Literature Review

Abstract

Social media data have been widely used in natural sciences and social sciences in the past 5 years, benefiting from the rapid development of deep learning frameworks and Web 2.0. Its advantages have gradually emerged in urban design, urban planning, landscape architecture design, sustainable tourism, and other disciplines. This study aims to obtain an overview of social media data in urban design and landscape research through literature reviews and bibliometric visualization as a comprehensive review article. The dataset consists of 1220 articles and reviews works from SSCI, SCIE, and A&HCI, based on the Web of Science core collection, respectively. The research progress and main development directions of location-based social media, text mining, and image vision are introduced. Moreover, we introduce Citespace, a computer-network-based bibliometric visualization, and discuss the timeline trends, hot burst keywords, and research articles with high co-citation scores based on Citespace. The Citespace bibliometric visualization tool facilitates is used to outline future trends in research. The literature review shows that the deep learning framework has great research potential for text emotional analysis, image classification, object detection, image segmentation, and the expression classification of social media data. The intersection of text, images, and metadata provides attractive opportunities as well.