Disruption or Adoption? Understanding Perceptions of Driverless Shuttles through Social Media Data Mining
Zheng, Max, Double Hoo Research Grant, University of Virginia
Mondschein, Andrew, Architecture - Department of Urban and Environmental Planning, University of Virginia
Jiang, Zhiqiu, Architecture Graduate, University of Virginia
Increased travel as a result of urbanization and population growth has led to the need for safer, more efficient transportation in US cities. We examines whether the public believes driverless transportation systems could meet this demand by analyzing public social media data from Twitter. Using mined Twitter data (2012 - present), we conduct a two-pronged approach to understand public perceptions of driverless technology. We used a Natural Language Processing approach using topic modeling to infer latent topics of interest related to driverless technology, and developed a sentiment analysis model to uncover the public dominant emotions towards each area of interest. Through topic modeling, we uncovered a set of 5 latent themes consisting of Safety Perceptions, Technology Development, Industrial / System Integration, Milestone and Vision, and Ethics and Policy as well as their classified opinions of positive and negative. The findings indicate that safety, technological progress, and industrial and urban integration are of major concern that largely affect the public’s acceptance of driverless technology. Driverless vehicle developers can leverage these results to influence what functionalities they should improve upon, and how they can shape their marketing campaigns to cater to customers’ needs and expectations. More importantly, the findings will help public transport operators and city planners as they attempt to integrate autonomous vehicles into the urban transportation system.
BA (Bachelor of Arts)
Machine Learning, Driverless, Data Mining, Twitter, Topic Modeling, Sentiment Analysis