Understanding Public Perspectives on Fracking in the United States using Social Media Big Data

Area of Science:

• Environmental Science
• Sociology
• Data Science

Background:

• Public attitudes towards hydraulic fracturing (fracking) are influenced by various factors.
• Traditional research methods like surveys have limitations due to small sample sizes and potential bias.
• A comprehensive understanding of public opinion on fracking is crucial for policy-making.

Purpose of the Study:

• To analyze national public attitudes toward fracking using social media data.
• To investigate the relationship between socio-demographic, economic, political, and environmental factors and public opposition to fracking.
• To identify spatial patterns and varying scales of these associations across the U.S.

Main Methods:

• Compiled geo-referenced social media data (Twitter) from 2018-2019 for the entire United States.
• Utilized multiscale geographically weighted regression (MGWR) to analyze county-level relationships.
• Examined correlations between negative fracking tweets and factors like income, race, education, unemployment, fracking sites, health insurance, and political affiliation.

Main Results:

• Socio-demographic factors like higher income, larger African American populations, and lower education levels correlate with less opposition to fracking.
• Regional disparities exist: higher unemployment, fewer local fracking sites, and greater health insurance enrollment are linked to increased opposition in specific U.S. regions.
• In the southern Great Plains, increased Republican voter share is associated with less vocalized opposition to fracking on Twitter.

Conclusions:

• Public attitudes toward fracking exhibit significant spatial heterogeneity and are influenced by a complex interplay of factors.
• Findings highlight geographical divides in how socioeconomic and political variables shape fracking opinions.
• The methodology provides a scalable approach for understanding public sentiment on controversial issues using big data.