Visualization Grammar: A Phrase-Based Reservoir Information Visualization System

Abstract Existing data query and visualization software capable of creating reservoir data visualizations are targeted at expert users and have a high learning curve. Also reservoir data are not usually universally accessible through a unified interface. In this paper we introduce a system called Visualization Grammar (VG) that allows users to generate compelling visualizations from large volumes of structured multi-dimensional reservoir data. VG can be easily deployed and requires virtually no train-ing from end-users. VG's goal is to create an intuitive and interactive visualization tool aimed at users of the petroleum industry with little or no expertise in data query and graphic design. These visualizations can in turn be used to analyze the ever growing volume of reservoir data. We describe VG's portable and intuitive user interface and its architecture and provide implementation details. Our results show how VG supports data access and abstraction, visualization query formulation and editing and automatic visua-lization generation. Phrases built by VG provide with a unified and universally intelligible representation that can support collaboration and help managing complex visualization requirements across domains of expertise. We show text, 2D and 3D visualization examples including volume renderings of reservoir data. Collaborative features further improve its usefulness for interdisciplinary data analysis. Consequently VG represents a unified data visualization interface that can be greatly bene-ficial to the petroleum industry. The VG system provides the following benefits. (1) The user interface is simple and intuitive requires virtually no training. It relies on a natural language formulation supported by a step-by-step visual guidance where the user is presented with valid data and visualization alternatives that are dynamically changing within the context. (2) A volume rendering ex-tension module with a GPU implementation provides interactive performance and improved perception. (3) VG's architecture is scalable and extensible providing access to a heterogeneous set of data sources. VG's visualizations therefore have the po-tential to greatly improve inter-domain collaboration and data analysis in the enterprise.