PiMi

Abstract

Monitoring indoor PM2.5 concentration and understanding the cause of indoor PM2.5 pollution are essential for people's health in urban areas. Previous studies under controlled experiments are hard to be generalized and datasets under real circumstances are with small scale limited by either the material resources or the manpower resources. To address these challenges, we utilize the methodology of participatory sensing to collect 133 million indoor PM2.5 samples along with labels of building characteristics and human behaviors, from 407 ordinary citizens in Beijing. Together with the simultaneous outdoor PM2.5 concentrations from national air quality monitoring stations, our dataset provide opportunities for studies on indoor air quality, time series analysis as well as time series modeling.