An Analysis on a YouTube-like UGC site with Enhanced Social Features

Abstract

YouTube-like User Generated Content (UGC) sites are nowadays entertaining over a billion people. Resource provision is essential for these giant UGC sites as they allow users to request videos from a potentially unlimited selection in an asynchronous fashion. Still, the UGC sites are seeking to create new viewing patterns and social interactions that would engage and attract more users and complicate the already rigorous resource provision problem. In this paper, we seek to combine these two tasks by leveraging social features to provide the reference for resource provision. To this end, we conduct an extensive measurement and analysis of BiliBili, a YouTube-like UGC site with enhanced social features including user following, chat replay, and virtual money donation. Based on datasets that capture the complete view of BiliBili---containing over 2 million videos and over 28 million users---we characterize its video repository and user activities, we demonstrate the positive reinforcement between on-line social behavior and upload behavior, we propose graph models that reveal user relationships and high-level social structures, and we successfully apply our findings to build machine-learnt classifiers to identify videos that will need priority in resource provision.