Graph modeling from relational databases

Abstract

The importance of mapping data from the relational model to other models and technology have been widely discussed. Bring them to graph model is an example of that discussion. The graph model allows to analyze the topology such as social networks, link prediction and recommendation systems. There are initiatives to map a relational database to a graph representation. However, they are not concerned with the different ways to generate a graph, especially when the main objective is to perform topological analysis. This work proposes some heuristics aiming a mapping systematization from relational model data to graph representation. The main contribution is to provide support for an adequate choice of the graph model, according to the analysis to be performed. Experiments are presented and show interesting results including heuristics to support the user on the graph model choice.