Research On Sentiment Inequality Of Residents In Nanjing Based On Weibo Check-in Data

Differences in the built environment of cities affect the quality of life of residents,which in turn can lead to differences in their experiences and perceptions,resulting in sentiment inequalities.Improving the quality of urban habitat through planning and design to promote the physical and mental health of residents has always been one of the goals of the urban and rural planning discipline.With the popularity of smart mobile devices and social media platforms in recent years,sentiments have become an important indicator of residents’ life experiences and mental health.Online tweets with location information can be used as objective information to reflect the sentiment differences of urban residents in different locations,overcoming the limitations of previous studies with small sample size or lack of spatial location information,and combined with deep learning-based sentiment quantification methods can identify the spatial pattern of urban residents’ sentiments.This paper uses sentiment analysis methods to quantify and analyze the sentiment information contained in 480,000 Weibo tweets within the central city of Nanjing in 2015 and draws a sentiment map to reveal the unequal spatial distribution of urban residents’ sentiments,and then explores the correlation between the distribution of land value at the macro level and the difference in built environment at the settlement level in the city and the unequal sentiment of residents.The results of the study indicate that:(1)The overall resident sentiment score of Nanjing’s central urban area is 0.643,indicating a positive overall sentiment.However,there are spatial differences in the distribution of residents’ sentiments in the downtown area of Nanjing,showing that the central area is better than the surrounding areas,and that positive sentiment is more concentrated and negative sentiment is more dispersed,with positive sentiment being more concentrated in the old downtown area,the west area of the river and the area around Baijia Lake,while negative sentiment is relatively dispersed in the non-core areas of Jiangbei,Jiangning and Xianlin regions and the urban fringe areas near the boundaries of the central city.(2)In the results of the study on the relationship between land value and residents’ sentiments,the concentration index(CI)in the overall level concentration curve is 0.083,indicating that residents in areas with high land value contribute more positive sentiment with the cumulative distribution of land value from low to high.The results of the study in different land use types show that residents in public administration and public service,residential,commercial and industrial land use by more susceptible to the influence of land value differences and exhibit sentiment inequality.There is no significant correlation between land value differences and residents’ sentiments inequality in other types of land use.(3)In the study of the relationship between built environment and residents’ sentiments in settlements,the relative importance of characteristics of neighborhood environment dimension,transportation condition dimension and community construction dimension accounted for 47%,34% and 17%,respectively,representing the difference in the degree of influence of different dimensions on residents’ sentiments.Combined with the analysis results of nonlinear regression,we were able to obtain different degrees of nonlinear influence relationships between different indicators and residents’ sentiments.The nonlinear thresholds for each indicator can provide a more accurate data reference for planning and construction standards.This paper uses social media tweet data to analyze the sentiment inequality of urban residents and explores the correlation between urban land value and residential built environment differences and residents’ sentiments.It can provide a reference for interdisciplinary research in public health,urban and rural planning,human geography and sociology.In practice,it provides powerful methodological and data support for planning decision makers to design targeted habitat interventions and provide a basis for promoting equity in the built environment of cities.