Asynchronous Digital Participation In Urban Design Processes: Qualitative Data Exploration And Analysis With Natural Language Processing
This paper aims to improve the usability of qualitative urban big data sources by utilizing Natural Language Processing (NLP) as a promising AI-based technique. In this research, we designed a digital participation experiment by deploying an open-source and customizable asynchronous participation tool, “Consul Project”, with 47 participants in the campus transformation process of the Singapore University of Technology and Design (SUTD). At the end of the data collection process with several debate topics and proposals, we analysed the qualitative data in entry scale, topic scale, and module scale. We investigated the impact of sentiment scores of each entry on the overall discussion and the sentiment scores of each introduction text on the ongoing discussions to trace the interaction and engagement. Furthermore, we used Latent Dirichlet Allocation (LDA) topic modelling to visualize the abstract topics that occurred in the participation experiment. The results revealed the links between different debates and proposals, which allow designers and decision makers to identify the most interacted arguments and engaging topics throughout participation processes. Eventually, this research presented the potentials of qualitative data while highlighting the necessity of adopting new methods and techniques, e.g., NLP, sentiment analysis, LDA topic modelling, to analyse and represent the collected qualitative data in asynchronous digital participation processes.
Keywords: Urban Design; Digital Participation; Qualitative Urban Data; Natural Language Processing (Nlp); Sentiment Analysis; Lda Topic Modelling; Sdg 10; Sdg 11.