A text mining-based thematic model for analyzing construction and demolition waste management studies

Over the years, numerous studies have been conducted to investigate construction and demolition waste (CDW) management problems. However, the massive amount of literature brings challenges to scholars because it is difficult and time-consuming to manually identify research emphasis from the literature. Therefore, a method that can informationize literature collection and automatically detect insights from the identified literature is worthy of exploration. This paper attempts to present a comprehensive thematic model by combining Latent Dirichlet Allocation, word2vec, and community detection algorithm on python to detect insights from CDW management literature. Based on the database of Web of Science, 641 articles published between 2000 and 2019 are retrieved and used as the sample for analysis. The comprehensive thematic results reveal a four-domain knowledge map in CDW management research, which covers (1) introducing current situation of CDW management, (2) quantifying CDW generation, (3) assessing CDW and by-products, and (4) facilitating waste diversion. Future research directions in CDW management research have also been discussed. The results prove that the comprehensive thematic model is useful in mining insights from CDW management literature.