Immersive And Collaborative Data Visualization Using Virtual Reality

2018 IEEE International Conference on Big Data (BigData 2018)

The key challenges of actionable knowledge discovery come from the data complexity, which often manifests as a high dimensionality of data spaces, where each measured quantity adds a new axis. The key bottleneck in the data analysis and understanding is an effective visualization; it is a bridge between the quantitative content of the data, and the human intuition and pattern recognition. Visualization is essential at every step: from data cleaning and preparation, the choice of appropriate analysis algorithms, the interpretation of their output, and the final presentation of the results. The premium is in visualizing effectively as many dimensions of the data space as possible: if there are structures in the data that involve multiple variables, e.g., clusters, correlations, gaps, outliers, etc., projecting them on a standard 2D plot hides or destroys such information. Virtual Reality (VR) offers a powerful new platform for an effective visualization of high-dimensionality data spaces, where the users (who can be geographically separated) can interact with the data, machine intelligence and analysis tools, and collaborate with each other in a shared virtual environment. Immersion amplifies the ability to grasp the relationships and patterns that may be present in the data, more effectively than in any traditional visualization platform, and even discover patterns that are impossible to see in any other way. We have developed an innovative data visualization and analytics platform that combines a multi-dimensional data visualization with a variety of Machine Learning tools, and that enables a collaborative data visualization in VR, using commodity VR headsets. We will illustrate its application on a variety of practical examples, and provide some comparisons of its effectiveness as compared to the traditional data visualization and analytics tools.