The dataset is the matrix named view_count in Youtube_f50_new.mat, where each row represents an hour and each column represents a file. The original data sets can be obtained from https://www.kaggle.com/nnqkfdjq/statistics-observation-of-random-youtube-video.