The autonomous vehicle social network: Analyzing tweets after a recent Tesla autopilot crash

Abstract

Automated vehicle technologies offer a potentially safer alternative than manually driven vehicles, but only if they are accepted and used appropriately. Social media platforms may offer an opportunity to assess peoples’ willingness to accept and use automated vehicle technology, but questions remain on the structure and content of the social media conversation. To answer these questions, we performed an analysis of tweets surrounding a recent Tesla Autopilot incident. Tweets were analyzed at three levels: term frequency, account tweet and retweet frequency, and sentiment. The most frequent terms of the conversation shifted from “amazon” and “start up” to “autopilot” and “vehicle” following the crash, however, the specific tweet contentreferenced an earlier event. A small portion of accounts were responsible for the majority of the tweets in the dataset, and were rarely ret weeted. Positive and negative sentiment decreased following the crash, suggesting that a more complex sentiment analysis is needed to gauge changes in public opinion of automated vehicles.