‘I Tweet about Our #GreenEnergy’—Automated Classification of Social Identity and Opinion Mining of the Dutch Twitter Discourse on Green-Energy Technologies

Abstract

Understanding the complexities of public opinion is crucial for a green-energy transition. This present study examines the sentiment of public opinion towards various energy technologies on Twitter during the Dutch 2021 general elections. A dataset comprising 186,822 tweets and profile descriptions was analyzed using two automated text classifiers to explore how individuals with different self-proclaimed identities perceive green-energy technologies. The analysis involved the application of the sentiment and social identity classifier models, followed by a frequency and co-occurrence analysis. The findings revealed a negative overall sentiment towards green-energy technologies in the Twitter discourse. It further showed that perceptions may differ depending on a technology’s development stage, with emerging technologies generally receiving more favorable views compared to established ones. Furthermore, it was found that, although there is a general trend of negative sentiment based on political identity, and positive sentiment based on occupational identity, this trend did not consistently apply to specific energy technologies. This discrepancy can likely be attributed to varying implementation effects and contextual situations associated with the technologies. The findings suggest that personalized communication strategies for specific social groups may be beneficial for understanding and addressing public opinions, needs, and concerns within the energy transition. The complexity of understanding public opinion in the context of green-energy highlights the need for a nuanced approach in future research.