Next-generation language-based data mining techniques for smart construction documentation: challenges and opportunities

Data mining (DM) has emerged as a crucial tool for analyzing large datasets, facilitating knowledge discovery, and informing data-driven decision-making. Since most data in the construction industry are in an unstructured textual format, DM techniques based on language understanding and text analysis can facilitate the utilization of diverse aspects of this data. Despite the increasing use of DM in the construction sector, no systematic classification has been attempted for language-based DM techniques and their future implications. This paper examines the literature on language-based DM techniques and classifies it into three major categories. A systematic review of 132 research articles in smart construction management, published in recent years (2021–2025), is conducted to showcase the applications and future potential of language-based DM techniques. This paper provides valuable insights into utilizing language-based DM techniques to enhance productivity, mitigate risks, and improve decision-making in construction management. The review identifies three dominant research streams—natural language processing, text mining, and large language model (LLM)-based methods—and highlights their most prevalent applications in construction, including contract analysis, building information modeling (BIM) integration, safety management, risk mitigation, and automated compliance checking. Key insights reveal rapid growth in LLM-driven workflows, persistent challenges with domain-specific datasets, and emerging opportunities for combining text-based analytics with BIM and smart construction technologies. This review provides academics and industry practitioners with a comprehensive understanding of the latest DM adoption in construction documentation.