As society progresses, air quality is becoming an increasing concern in people’s lives. We selected hourly recorded data from Lanzhou, China, between 13 May 2014 and 18 April 2020 as the training and test sets. The data itself was first explored by exploratory analysis and pre-processing, exploring the data itself through visualisation and other means and selecting appropriate feature values to put into the subsequent training. A random forest classification model was then built using pyspark and MLlib packages, and the model was trained and evaluated. After finding the optimal combination of parameters through parameter search, the model eventually made classification predictions for this dataset with an accuracy of 0.64. The experimental results show that the random forest classification model has good performance in predicting the air quality in Lanzhou city and has good application prospects.
