Large-scale graph processing is an important rapidly-growing area of development because many real-world big-data applications are intuitively represented as graphs (e.g., the internet, the web, social networks, etc.)  Due to the complexity of graph algorithms and the need for rapid algorithm development and evolution, designing a graph processing framework which is both easy to use and has good performance is challenging.

In this project, we explore custom graph analysis algorithms on large-scale data set. We conduct this study with a focus on new graph algorithms for link prediction, community detection, and graph sparsification algorithms. For this exploration, we make use of the Oracle PGX framework, exploiting Green-Marl DSL feature for implementing custom algorithms. The usability and performance of PGX would be investigated by comparing against other frameworks such as Apache Giraph and Apache GraphX.