Dissecting Electric Vehicle Conversations on Twitter/X: A Tree-Based Analysis of Discussions

Abstract

Electric vehicles (EVs) have become a focal point in public debates on climate change and transportation, particularly in Europe following the European Parliament's decision to ban the sale of new combustion engine vehicles by 2035. This study analyzes French conversations on EVs using a novel open-source dataset of Twitter discussion trees. The dataset includes over 1 million tweets, structured as 22,635 discussion trees, and offers insights into the dynamics of online discourse about EVs. We manually annotated a subset of tweets and fine-tuned a BERTweetFr model for automatic classification and sentiment analysis. Our findings reveal specific structural patterns in online discussions, varied behavior and influence of key actors, and thematic contexts in which EV-related topics are addressed. This research contributes to understanding public opinion on EVs and provides a methodological framework for analyzing online conversations.