Modeling Urban Air Quality Trend Surface Using Social Media Data

Air pollution is getting worse with the development of cities in recent years. Urban air quality is mainly monitored by air quality monitoring stations at present. However, the number of stations is limited and the air quality fluctuates in different urban areas. So it is unefficient to detect air quality's distribution in a city by air quality monitoring stations only. Based on Sina Weibo data with location information, we propose an urban air quality trend surface modeling method by analysing the correlation between air pollution related topic microblogs and air quality monitoring station AQI data. The study reveals that our method not only qualitatively shows the relative air quality in diffferent regions of the city, but also demonstrations the urban air quality in a quantitative and fine-grained way. The findings of this study evaluate the feasibility of using a new type of large-scale data source for research on air quality estimation of any location in a city, and are of great significance when reflecting air quality distribution and finding areas where are relatively air polluted.