Sentiment variations affected by urban temperature and landscape across China
Hongbin He, Ranhao Sun
Abstract
Understanding the ramifications of urban environments on human well-being is of increasing importance. However, extant studies often focus on specific cities, neglecting the broader national framework. Here, big data techniques were utilized to analyze >6.5 million geotagged Weibo from China in 2021. Each Weibo's sentiment was calculated using advanced deep learning-based sentiment analysis, and the effects of temperature and green spaces on sentiments across different urban magnitudes and genders were investigated. The results revealed significant sentiment variations among different urban magnitudes, with males exhibiting less positive sentiments than females. Additionally, a reverse 'U'-shaped distribution was observed between sentiments and urban temperature across all urban, with more positive sentiments occurring within the temperature 15 °C to 25 °C. Notably, urban green spaces play a crucial role in enhancing sentiments across different urban magnitudes. Nonetheless, the analysis indicates that sentiments are not solely influenced by urban magnitudes, temperature, and the normalized difference vegetation index (NDVI) individually, but also by interactive effect. The insights gleaned from this study can contribute to formulating urban policies and measures aimed at optimizing the settlement environment, fostering positive sentiment, and thereby ameliorating overall well-being.