Automatic Transformation of Data Warehouse Schema to NoSQL Graph DataBase: A Comparative Study

Driven by the continuous growth of data generated by social networks (SN), traditional data warehouse (DW) approaches need to evolve. Commonly used logical models such as star, snowflake, or constellation schemas prove to be insufficient when handling social data, which demands more scalable and flexible systems. As a result, NoSQL systems are emerging as a promising alternative. In the context of Big Data warehouses, a graph-oriented NoSQL database system is considered an ideal storage model due to its strong suitability for data warehousing and online analysis. The use of NoSQL models facilitates data scalability, while the graph store provides greater flexibility in storing and managing large volumes of data. In the absence of a clear approach which allows the implementation of data warehouse under NoSQL model, in this paper we propoe new rules for transforming a multidimensional conceptual model into NoSQL Graph oriented model. We distinguish two types of transformation: a simple one and a hierarchical one. To validate our transformation rules, we implemented two data warehouses using Neo4j and Java routines in Talend Data Integration tool. These systems were evaluated in terms of ”Write Request Latency” and ”Read Request Latency” using LDBC-SNB benchmark.