U.S. and China Flight Delay Datasets(including delay, weather categories, airport features, and aviation network crowdedness matrices)

The U.S. delay dataset is collected from Kaggle(https://www.kaggle.com/datasets/robikscube/flight-delay-dataset-20182022), covering three years of flight data from January 1, 2017, to December 31, 2019. The dataset originally collected includes data from 360 airports. We remove airports with fewer annual flight numbers and select data from 75 medium and large airports for our experiments. The weather dataset summarizes 12 weather categories, including normal weather, light rain, moderate rain, heavy rain, light snow, heavy snow, moderate fog, severe fog, precipitation, storm, hail, and severe cold. The China delay dataset was collected from Xiecheng(https://pan.baidu.com/s/1dEPyMGh\#list/path=\%2F), covering two years of flight data from May 1, 2015, to June 1, 2017. We select airports with a total number of flights exceeding 10000. From related special event data, 10 weather categories are obtained, including normal weather, thunderstorms, cloud, thunder, fog, strong winds, storms, snow, and severe convective weather. During the experiment, we only consider flight records between 6:00 AM and 11:59 PM, as very few flights were observed outside this time frame.