User Behavior Analysis of a Mobile Short Video Application

Video accounts for a high proportion of Internet traffic and this proportion will continue to rise in the future. The mobile short video application represented by TikTok has a large user scale and is growing rapidly. An in-depth understanding of user behavior in mobile video systems can help us optimize video content distribution to reduce network pressure. In this paper, we have analyzed the user viewing data of mobile short videos and compare the results with traditional video analysis results. We found some characteristics of users' online mode, complete viewing, and likes. We also found that the popularity ranking and the like ranking of short video data conform to the Zipf-like formula. Finally, we did an experiment to prove that our data simulation system based on the Zipf-like formula can work well on short video data.