Understanding economic and environmental tradeoffs of bottled water facilities using structural topic modeling and Lexicon-based categorization of public news media

Abstract Bottled water facilities exist across the United States (U.S.) in all 50 states and have the potential to affect localities in which they are located. This study aims to understand how water bottling facilities are portrayed in news media in the U.S., focusing on economic and environmental tradeoffs, by using Natural Language Processing techniques, specifically Structural Topic Modeling and Lexicon-based Categorization, across different U.S. states and time periods. Through our stratified analysis, we identified key environmental topics and natural resources, as well as companies attracting media attention in different regions and time periods. Results suggest that: (1) the increase in news media publications were correlated with current events such as drought or the start or change in operations of bottling facilities, and (2) these current events also influenced whether the coverage focused on economic topics or environmental concerns. The balance of water availability and economic development is a theme prevalent among the results of both forms of analysis. This study demonstrates the importance of understanding the unique values of a locality before making decisions that may affect residents.