China's urban air quality evaluation with streaming data: A DEA window analysis

Abstract

After the promotion of urbanization in the past decades, air pollution has become one of the bottlenecks of China's urban development. Kinds of real-time air pollution indicators are recorded by using environmental detection system in China's urban area that produces precious streaming data. The present paper constructs window DEA model to compute the dynamic air quality index after applying hierarchy analysis to resolve the heterogeneity of time varying data. In the section of empirical study, we select the daily data of CO, NO, SO2, PM 2.5 and PM 10 since January 2018 to August 2019 for 360 cities in China which includes 1,092,600 streaming data. Our empirical findings indicate that air pollution is heavily serious in most China' cities, in which more than 95% cities have over emitted air pollution by 30% at least. Chinese urban air quality is significantly affected by the change of month and shows an inverse U-shaped curve relationship in year, while the orders of weeks within month or order of days within week is irrelevant. The provinces with the most urban air pollution concentrate in middle China and from a continuous pollution zone with Shanxi as the center.