A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Information and Communications Technologies, School of Engineering and Technology

Thesis (M.Eng.) - Asian Institute of Technology, 2009

A data warehouse is a huge collection of data designed specifically for reporting and analysis. Online Analytical Processing (OLAP) tools helps users to analyze, explore and provide answers to analytical questions for multidimensional data in a data warehouse, often using different visualizing techniques. Combination and coordination between multiple visualizations provides more information to detect relations and dependencies than a single visualization. One example of coordination between views is 2brushing,3 which is simultaneous highlighting of the data of interest in the coordinated views. Coordinated visualization tools are often developed using either a data-centric coordination architecture, which requires complex joins between views, or a polling architecture, which is ineffecient. To solve these problems, this thesis proposes a brushing coordination architecture based on pushing changes from the model out to the relevant views. A prototype was developed over a datawarehouse using an open source toolkit named JRubik. Change in the OLAP model will be published to the other views, solving the problem of constant polling to the coordination component. The architecture was evaluated on performance metrics, and the results showed that coordinated brushing based on this architecture could be added with little effort or impact on the performance of the application. Furthermore, the evaluation was done based on the extendibility and reusability issues for other views, which showed, the views can be added to JRubik with slight modification. To conclude, a brushing coordination architecture for multiple views was developed based on an OLAP model and brushing component that provides flexibility in brushing between different independent pluggable visualization.