| description abstract | The coronavirus disease 2019 (COVID-19) pandemic has brought unprecedented impacts (e.g., labor shortage, suspension and cancellation of projects, and disrupted supply and logistics) on the US construction industry. To address challenges caused by the pandemic, it is critical for the construction industry to develop a clear understanding of how the pandemic has affected the industry and how it will change in the future. However, assessing the impacts of COVID-19 on the construction industry is challenging due to the broad influence of the pandemic and the dynamic nature of the industry. The Purdue Index for Construction (Pi-C), which was developed as an indicator based on five dimensions and corresponding metrics to measure the health status of the construction industry, offers an opportunity to understand the impact of the pandemic. In this context, this paper presents a study to reveal the relationship between COVID-19 and the health status of the industry as measured through Pi-C and predict the future trend of the construction industry. This study achieves the objective via the three steps. First, the relationship between the pandemic and Pi-C metrics is identified using the Granger causality test and structural equation modeling (SEM) analysis. Second, multivariable prediction models are developed based on a long short-term memory (LSTM) network—a deep learning algorithm—to predict Pi-C metrics in the future. Third, forecasted Pi-C metrics are integrated into the existing Pi-C structure to analyze the impacts of the COVID-19 pandemic and predict its trends in 2021–2022. The results revealed that the impacts of the pandemic were conspicuous in two Pi-C dimensions (economy and stability), whereas no significant impacts were observed in the remaining Pi-C dimension (social). In addition, the Pi-C forecasted that there would be no significant adverse impacts on the US construction industry caused by the pandemic until the end of 2022. | |
