Air Pollution Perception Bias Based on Social Media: Spatiotemporal Pattern & Mechanism

1. Introduction

 The project aims to measure the public’s perception of air pollution using geo-tagged weibo data (the Chinese Twitter) and to further discuss the discrepancy between perceived and actual air pollution levels and the rationale behind it. We used the web crawler technology to obtain the daily monitoring data of China’s air monitoring sites from 2012 to 2014 and tens of millions Weibo records in the corresponding period. Taking the data when the air quality is good as the benchmark, we represent the public’s perception of air pollution through the fluctuation of weibo check-in number and weibo emotional fluctuation, moreover, the air pollution perception deviation was measured by the difference between the normalized perception index and AQI.