Designing Participatory Landscapes: Integrating AI-Based Sentiment Analysis into Community-Involved Landscape Planning

This study develops an AI-enhanced framework to address limitations in capturing nuanced community sentiments for landscape planning. The proposed methodology integrates natural language processing (customized BERT model) with geographic information systems through a three-stage analytical pipeline, enabling fine-grained sentiment classification and spatial mapping of unstructured public feedback from digital platforms. Validated through urban waterfront regeneration, rural heritage conservation, and post-disaster reconstruction case studies, the framework demonstrates superior performance compared to traditional methods: greater input coverage, 58% faster processing, and 68% adoption rate of sentiment-informed design modifications (versus 41% baseline). The system uniquely bridges qualitative public sentiments with quantitative spatial planning parameters through geo-referenced text analysis. The research makes dual contributions: (1) advancing participatory planning practice by revealing latent design priorities through scalable sentiment analysis, and (2) enriching human-computer interaction research with a replicable model for contextualizing subjective feedback in socio-cultural landscapes. By maintaining analytical rigor while accommodating diverse digital participation channels, the framework democratizes community engagement without sacrificing sensitivity to local contexts. These innovations establish new standards for landscape design that is both data-driven and culturally responsive, particularly valuable for complex projects requiring balanced integration of technical and social dimensions.