Formalizing Graph Database and Graph Warehouse for On-Line Analytical Processing in Social Networks

Abstract

Graph databases have been widely employed for representing connected pieces of information for different kind of domains. The data model embraces relationships as a core aspect to connect objects, and organizes everything into a network for efficient query processing of versatile applications, e.g., on-line social networking, metropolitan traffic modeling, marketing channels simulations or even counterterrorism analysis. As an emerging technology for encoding network structures, graph databases are also widely used as an infrastructure for social network analytics, which help us understand some phenomenon or hidden knowledge in buzz marketing, technology trends or public issues regarding social behaviors. Although many graph database management systems have been developed, there are still no formal definitions for theoretical graph database modeling. In this paper, we will present a formal definition for graph database model, extend the concept of data warehouse into graph warehouse, and define the basic elements of a graph warehouse for the development and derivation of graph-based multi-dimensional business intelligence through on-line analytical processing (OLAP) on graph databases.