A Framework for Environmental Policy Social Analysis: Quantifying Social Impact Through Guided and Unguided Data Analyses

Abstract

In environmental policy analysis, there is a general lack of a systematic way in evaluating the social impact of a policy. This paper seeks to fill that gap by presenting a framework to analyze how environmental policies interact with the public masses. This paper looks at the Paris Climate Agreement and the Green New Deal using the proposed framework, which includes guided data analyses such as survey analyses and unguided data analyses such as sentiment analysis and topic modelling on social media data. While this paper has more of a focus on unguided data analyses, it has shown that survey analyses allow researchers to not only quantify the support of a policy at any given moment but also allow them to the pinpoint variables that might affect the level of support for that policy. For unguided data analyses, this paper uses three different packages for sentiment analyses on twitter data and constructs three different topic modelling models on the same data. Afterward, sentiment analysis is performed again on the resulting topics. Overall, the Flair package appears to be the best at identifying the correct sentiments of the environmental policy tweets while the non-negative matrix factorization model with Frobenius norm is best at identifying the topics of most discussion. In conclusion, both guided and unguided data analyses have their merit and the wealth of information that can be obtained from either should not be ignored. This framework can potentially offer a way for policy analysts to provide meaningful and comprehensive results for policy makers and enforcers to use in policy making/adjustment.