Analysis of Influential Factors in Secondary PM2.5 by K-Medoids and Correlation Coefficient

There are many influential factors in PM2.5, reducing the emission of PM2.5 is one of international subjects. In recent years, it is indicated that one of the sources of secondary PM2.5 is the complex chemical reaction between NH₃ and air pollutants (VOCs, particulate matter, NO_x, SO_x). The Committee on Agriculture of FAO indicates that 64% of NH₃ emission on the earth surface is derived from stock raising which motivates this study to discuss following two subjects based on Open Government Data. Subject 1 calculates the effect of the controlled air pollutants (VOCs, particulate matter, NO_x, SO_x) and the quantity of livestock (e.g. pigs, chickens and so on) nearby the air monitoring stations on the annual mean of PM2.5. Subject 2 uses Apache Spark as Cloud computing platform, the air monitoring stations are geographically clustered by K-medoids to calculate the Spearman's correlation coefficient of pollution source and PM2.5 of each cluster. The experimental results show that the monitoring station with more air pollutants and livestock raised nearby has higher annual mean PM2.5 concentration. The results are expected to provide the government bodies to make environmental decisions and the plants and livestock farms to install air monitors to analyze the air quality data. Our ultimate goals are to improve the environment and reduce both the emission of PM2.5 and the probability of getting cardiovascular disease.