SIGNATURE WORK
CONFERENCE & EXHIBITION 2022

With prevalent “stay at home” orders implemented in the United States since the breakout of the COVID-19 pandemic, human mobility was greatly limited. Under lockdown, urban green space (UGS) as an important social capital plays a crucial role in mediating people’s mental states. By affecting emotion towards the pandemic, differences in access to urban parks can lead to polarization in policy preference among residential groups. Yet few studies have examined UGS’s influence on public opinion on Covid-19. Therefore, this study uses Twitter discussion on COVID-19 and visit patterns to natural parks released by SafeGraph to bridge this literature gap. I labeled the theme of each tweet and their associated sentiment attributes with a sentiment analysis model based on XLNet and a Latent Dirichlet Analysis topic modeling model. Using a time-lagged fixed-effect regression model, I show that visit to natural parks is negatively associated with the expression of anxiety and pessimism and positively associated with the expression of optimism sentiment. The correlation is significant even when controlling the education level of each state. This finding shed light on how the natural environment may enhance opinion polarization among segregated urban residential groups.