Using ChatGPT to create a renewable energy sentiment index for Australia

Climate change has heightened the importance of understanding public sentiment towards renewable energy because it influences policy and investment decisions. This study develops a renewable energy sentiment index for Australia by examining tweets related to renewable energy. Its sentiment analysis approaches include three lexicon-based methods and ChatGPT, and they are compared using Plutchik’s three-dimensional circumplex model. The research employs business analytics techniques such as web crawling, text mining, and data mining to measure and analyse opinions, emotions, and sentiments in written documents. The study aims to provide insights into the factors that shape public perceptions of renewable energy and to assess the effectiveness of various sentiment analysis approaches, while also offering valuable information for policymakers and stakeholders in the renewable energy sector.

ChatGPT is applied to sentiment analysis in this study, which compares the performance of lexicon-based approaches with that of ChatGPT to evaluate their effectiveness in extracting sentiment information from social media content. The study finds that ChatGPT provides significantly higher agreement with human-based polarity classification of tweets than lexicon-based approaches provide. Overall, our findings suggest that ChatGPT seems to provide results superior to those of lexicon-based sentiment classification approaches. Furthermore, our results reveal that the level of joy expressed in renewable energy tweets decreased from 59.5% to 54.1% between the two time periods. Moreover, there was an increase in the level of anger expressed in tweets, from 34.8% to 40.1%, following the Russian attack on Ukraine. Thus, our findings suggest that overall sentiment expressed on Twitter regarding renewable energy in Australia has become less positive and more negative since the war in Ukraine began. It is important to note that the study did not investigate the causes behind the observed changes in sentiment. Nonetheless, our findings highlight the need to monitor changes in public sentiment towards renewable energy and to identify factors that could potentially contribute to shifts in sentiment.