Finding academic concerns of the Three Gorges Project based on a topic modeling approach

Abstract

The Three Gorges Project (TGP) has gone into the overall completion acceptance stage in 2014. As the world's largest hydropower project, the TGP has attracted worldwide attention over the past few decades. Previous studies mainly focused on a single aspect, such as engineering technologies, social impacts and environmental impacts, of the TGP. However, a large-scale review gathering systematic data to find academic concerns about the TGP is missing. Topic model is a text mining approach for discovering latent topics in a collection of documents. In this article, an emerging topic modeling approach, Latent Dirichlet Allocation (LDA), was introduced to uncover the intellectual structure of the academic literature focusing on the TGP. A collection of 8280 Chinese research articles highly related to the TGP was established with a time frame ranging from 2001 to 2013, and an 18-topic model was used to describe the intellectual structure. Two novel bibliometric indicators, including topic proportion and topic trend, were constructed to describe the academic concerns of the TGP. Topic proportion analysis shows that post-construction issues, including the social and environmental impacts brought by the TGP, have attracted more attention than the construction issues. “Ecology”, “Reservoir Operation”, “Land Administration”, and “Water Pollution”, have become the dominant research topics regarding the TGP during these years. Meanwhile, “Construction Technology” and “Design”, have gradually lost scholars’ interest. The results show that the approach reported in this study can provide sound and credible conclusions of the major academic concerns for a hydropower project. The topic modeling approach is expected to be widely applied as a methodological strategy in future hydropower and other infrastructure project assessment.