Streaming Graph Analytics for Real-world problems

Understanding dynamic graphs is essential to relevant analysis of near real-time systems, such as a case study we will present on tracking population-scale phenomena. Modeling the evolution of the graph explicitly leads to computational challenges. Hence, the innovation in our approach is using graph statistics in lieu of explicit modeling. Our streaming graph analytics use two phases; in the first phase, we track vertex and/or edge statistics using STINGER, a parallel graph data representation, and in the second phase, we generate big data insights.