Big Log Analysis for E-Learning Ecosystem

Recently, e-Learning emerges a rapid development, and e-Learning ecosystem has been proposed to provide sustained and stable services to cope with the growing demands of e-Learning. In e-Learning ecosystem, the amount of e-Learning log grows exponentially, introducing the variety and complexity of e-Learning log analysis. Therefore, a robust, scalable and practical logging architecture is urgently needed. Firstly, the characteristics of e-Learning log and its analysis are studied in this paper. Specifically, e-Learning log implies complicated characteristics, such as multi-dimensional correlation, heterogeneous multi-source, and cascading generation. Furthermore, the log analysis represents abundant diversity of demands, variety of methods and low-latency requirement in computation. Thereupon, this study presents a comprehensive logging architecture covering the whole life cycle of e-Learning log data which includes log collection, transport, storage, computation and service. To verify the proposed logging architecture, a related experimental implementation is developed for a realistic e-Learning ecosystem, and three typical e-Learning analyses are proposed.