How urban environment shapes EV charging experience in Travis County, Texas

Abstract

As electric vehicle charging stations (EVCSs) become increasingly embedded in cities, understanding user experience is critical for designing infrastructure that is functional, accessible, and comfortable. This study integrates AI-driven sentiment analysis with spatial modeling to examine how urban environments shape EVCS perception. Using more than 4,000 user reviews from Travis County, Texas, three large language models classified sentiment across categories such as charging operation, accessibility, and parking. Random Forest regression results show that walkability, greenery, openness, and surrounding amenities are among the strongest predictors of user sentiment, while operational concerns are widely distributed in peripheral areas and accessibility and parking frustrations cluster in dense commercial zones. These findings demonstrate that EVCS usability is influenced not only by technical performance but also by contextual qualities of the built environment. This work provides a scalable framework for real-time monitoring of EVCS experiences and actionable insights for location-sensitive planning that supports user wellbeing and sustainable mobility adoption.