PiMi is a participatory indoor air quality data sharing project launched in January 2014. Hundreds of PiMi air boxes, a low-cost indoor air quality monitor, were given out to volunteer users across China. The volunteer users access the PiMi air box for his personal air quality data through smartphones, and then upload the data to the PiMi cloud for further analysis. Having accumulated large amount of indoor air quality data under different circumstances, the PiMi cloud is able to apply statistical learning methods to model the relationship between indoor and outdoor air quality, and detect point of interests (POIs) in the data series. Together with the user-reported physicality information on the indoor environments, we are hoping to quantitatively evaluate the impacts of the environment physicality and human behaviors on the indoor air quality, and ultimately on how to alleviate indoor air pollution.