Graph BI & Analytics: Current State and Future Challenges Amine Ghrab 1,2 , Oscar Romero 3 , Salim Jouili 1 , and Sabri Skhiri 1 1 EURA NOVA R&D, Mont-Saint-Guibert, Belgium firstname.lastname@euranova.eu 2 Universitat Polit` ecnica de Catalunya, Spain oromero@essi.upc.edu Abstract. In an increasingly competitive market, making well-informed decisions requires the analysis of a wide range of heterogeneous, large and complex data. This paper focuses on the emerging field of graph ware- housing. Graphs are widespread structures that yield a great expressive power. They are used for modeling highly complex and interconnected domains, and efficiently solving emerging big data application. This pa- per presents the current status and open challenges of graph BI and ana- lytics, and motivates the need for new warehousing frameworks aware of the topological nature of graphs. We survey the topics of graph model- ing, management, processing and analysis in graph warehouses. Then we conclude by discussing future research directions and positioning them within a unified architecture of a graph BI and analytics framework. 1 Introduction Graphs are fundamental and widespread structures that provide an intuitive abstraction for the modeling and analysis of complex, heterogeneous and highly interconnected data. They have the benefit of revealing valuable insights from content-based and topological properties of data. The great expressive power of graphs, along with their solid mathematical background, encourages their use for modeling domains having complex structural relationships. In the context of Big Data, the focus of organizations is often on handling the rising volume of their data. However the variety and complexity of data through the different phases of data capturing, modeling and analysis is at least equally important. The variety challenge is the most critical challenge in big data nowadays, and efficiently handling the variety of data sources is considered to be the main driver of success for data-driven organizations [1]. Graphs meet the requirements to be the perfect canonical data model for data integration systems [2] given (1) their capability to deal with semantic relativism and semantic heterogeneities, (2) they are semantically richer at least as any other model (so they can represent any semantics), (3) they allow to create multiple views from the same source, (4) and most importantly, graphs are extremely flexible to compose new graphs. That is given two graphs, with one single edge a new graph could be directly created Ghrab, A. [et al.]. Graph BI & analytics: current state and future challenges. A: International Conference on Big Data Analytics and Knowledge Discovery. "Big Data Analytics and Knowledge Discovery, 20th International Conference, DaWaK 2018: Regensburg, Germany, September 3-6, 2018: proceedings". Berlín: Springer, 2018, p. 3-18. The final authenticated version is available online at https://doi.org/10.1007/978-3-319-98539-8_1