DEVELOPMENT OF TOOLS TO GENERATE GRAPH CUBE BASED ON MULTIDIMENSIONAL NETWORK ON GRAPH DATABASE

• Main Author: Djauhari Djohan, Albertus
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/43023
• Institution: Institut Teknologi Bandung

The amount of graph data is increasing with the increase of popularity of web based application, especially social media. Web application like social media produces large amount of graph data. Analysis of this data is not suitable using traditional data warehouse. The more suitable model for this problem is graph cube. Graph cube is set of aggregate networks, and aggregate network is graph from aggregation of multidimensional network based on certain set of attributes using certain aggregate function. Multidimensional network is a network modeling into vertices with multiple attributes and edges which represent relation between those vertices. Because there is no tool to generate graph cube, a solution to build a tool to generate graph cube is proposed. The development of the tool is started by describing tool’s requirements. From this description, functional requirements is created. Use case of the tool is created based on the functional requirements. Then, all modules needed is listed. This tool can generate graph cube from user’s input. The graph data for the input must be homogenous graph, which means it only has one type of node. The graph data can has several types of edge, it can be undirected or directed, and it can has hierarchy in the node’s attributes. From the generated graph cube, tool can display a graph representing an aggregate network based on selected dimensions and attributes. The tool has satisfied defined functional requirements. Evaluation shows that graph cube can reduce response time of given queries.