Assessing the Effects of Electric Vehicle Adoption on Urban Energy Structure Transition: A Geospatial Machine Learning Study in Beijing

Electric vehicles (EVs) have been proposed as a key solution for decarbonizing urban transportation and addressing climate change. As the use of EVs increases in cities worldwide, it may lead to significant transformation in urban development, including changes in the electrical system and people's travel behavior, such as charging preferences and choices of where to live and work. Some questions arise, will the rise of EVs lead to more suburbanization or drive people towards a more compact urban form? Additionally, how can the relationship between EV users' residential locations and new energy infrastructure be best coordinated? A study in the rapidly growing metropolis of Beijing aims to address these questions by combining geo-spatial big data analysis, machine learning, and theories of urban development to understand the relationship between EV users' residential locations and new energy infrastructure. A novel data mining strategy was proposed to identify actual EV users based on location data from smartphones. By analyzing observation data of EV users, the study applies the Gradient Boost Decision Tree model to examine the nonlinear associations between the spatial distribution of EV residents and neighborhood attributes such as employment density, GDP, land use mix, public charging accessibility, building areas, access to public transit, and suburbanization. The results indicate that a higher percentage of EV users prefer to live in areas that are neither too far away from the city center nor too close to it, particularly the threshold effects show that they are concentrated in areas where it has a 10 km distance from the city center. Additionally, the study found that most public charging activities tend to occur within 1.5 km from home, suggesting an optimal threshold for public charging station deployment. The findings of this study can help inform energy management and infrastructure planning at the local, regional, and national levels to promote sustainable urbanization and smarter energy planning in policy-making.