EAGER: Large-Scale Real-Time Information Visualization on Immersive Platforms

这个提议在快速发展的沉浸式分析领域开辟了新的领域。虚拟现实和增强现实技术非常适合于实现大规模数据分析，提供数据景观的全球概述以及细粒度控制和导航。拟议的工作将实现一个可扩展的开发环境，用于沉浸式分析体验，并评估使用虚拟现实耳机，显示墙和弯曲沉浸式显示器创作和协作探索非常大的数据集的不同方法。这项提议的工作将为让大量受众更容易获得沉浸式大规模数据探索铺平道路，开辟了让家庭虚拟现实用户参与新颖学习体验的可能性。PI将在定期的外展活动中展示由此产生的沉浸式可视化，以吸引未被充分代表的少数民族参与新的人机交互，从而促进STEM研究。有效的信息可视化需要能够实时处理和显示大量数据，以实现交互式分析。PI将在现代信息可视化库设计中构建直接图形处理单元（GPU）支持，目标是与当前基于web的可视化库相比，在渲染大量图形元素时加速一到两个数量级。这种计算的可扩展性将支持对非常大的数据集进行交互式探索，可以使用单一的可视化范式（例如节点链接网络）或多种范式（例如地图、网络、传感器数据、时间序列等）进行显示。沉浸式信息环境有可能支持比桌面或基于web的平台更有效的协作分析能力，通过将利益相关者群体置于大规模数据可视化的中间，以与他人的第一人称视角相关联，并交互式地共同导航信息空间，控制和突出不同的方面。PI将设计专注于这一潜力的交互式协作数据导航和控制工具。

This proposal breaks new ground in the rapidly evolving field of Immersive Analytics. Virtual and Augmented Reality technologies are uniquely suited to enable large-scale data analysis that provides global overviews of data landscapes as well as fine-grain control and navigation. The proposed work will implement a scalable development environment for immersive analytics experiences, and evaluatedifferent methods for authoring and collaboratively exploring very large datasets using Virtual Reality headsets, display walls, and curved immersive displays. The proposed work will pave the way for making immersive large-scale data exploration more accessible to large audiences, opening up possibilities of letting home virtual reality users participate in novel learning experiences. The PI will showcase the resulting immersive visualizations at regular outreach events to attract underrepresented minorities to novel human-computer interaction, and thus STEM research.Effective information visualization needs to be able to process and display large amounts of data in real-time to enable interactive analysis. The PI will build direct graphics processing unit (GPU) support into a modern information visualization library design, targeting one to two orders of magnitude acceleration in the rendering of large numbers of graphical elements, compared to current web-based visualization libraries. This computational scalability will enable interactive exploration of very large datasets, either displayed using a single visualization paradigm (e.g. node-link network) or multiple paradigms (e.g., map, network, sensor data, time series, etc.). Immersive information environments have the potential to support more effective capabilities for collaborative analysis than is possible with desktop or web-based platforms by placing groups of stakeholders in the middle of large-scale data visualizations to relate to others' first-person perspectives and interactively and jointly navigate through information spaces and control and highlight different aspects. The PI will design interactive collaborative data navigation and control tools that focus on this potential.