Knowledge is power – but extracting a useful knowledge from the data is not always easy.
We live at an extraordinary time. Information technology is transforming essentially every aspect of modern society: how we work, play, learn, and communicate. The pace of the transformation is unprecedented with the major changes happening on the scale of years rather than decades or centuries. One consequence of this revolution is an exponential growth of data rates and data volumes, reflecting Moore’s law that describes the rapidly evolving technology that generates the data. Just as important is the growth of data quality and data complexity.
Whether we are studying stars, hospital patients, or exchange-traded funds, for each object of the study we are measuring many quantities. Each of them is a column in a large spreadsheet, or a data dimension in a technical parlance. Together, they make a highly dimensional, abstract data space which may have tens, hundreds, or even thousands of dimensions. There may be complex patterns that encode the hidden knowledge present in the data: multivariate correlations, clusters, outliers/anomalies, gaps, etc. There can be an interplay of multiple factors – for instance, the gain in the market value of a given stock may depend on a complex combination of the economic and marked indicators but be different for the stocks in different sectors or different geographical regions.
“Big data” is not about the data. It is about discovering knowledge and insights in the data. Data are of no use if we cannot extract a useful, actionable knowledge from them and do so quickly and efficiently. Powerful statistical and machine learning tools do exist, but their applicability for a given problem depends on the overall distribution and geometry of data in a highly dimensional feature space.
Visualization is the bridge between the quantitative information in the data and the human intuition and understanding. We have a powerful pattern recognition system in our heads and we feed it mainly through vision. We have to look at the data in order to choose the right method for their analysis and to interpret the results. Yet typically we look at the data dimensions (e.g., columns in a spreadsheet) one at a time (histograms, pie charts) or two at a time (lines or scatter plots in 2 dimensions). Traditional visualization, confined to a flat screen or a flat paper, is inherently limited. Complex structures that may exist in 3 or more dimensions are lost when projected to a lower dimensionality display.
This is where Virtual Reality comes in. It does a lot more than just adding one extra dimension to a data display. When displayed in a 3D space that can be easily navigated, several more data dimensions can be encoded – and understood – through the use of colors, shapes, transparencies, animations, etc.
But wait! That’s not all! It turns out that a human perception of the patterns and relations present in the data works much better when we are fully immersed in such a data space, looking at the data from the inside out rather than from the outside looking in, as is the case in all traditional visualization approaches. This effect has been demonstrated by a number of research studies in different domains. We are creatures optimized to deal with the physical 3D world in which we are immersed, and our minds are most effective if we are looking for patterns in such a space even if the space itself is abstract in nature.
Furthermore, VR is a natural platform for collaborative data visualization and visual exploration. Users can interact with the data and with each other in a shared virtual space even if they are continents apart in the physical world. Such interactions are vastly better than any teleconferencing experience. They are already almost as good as physically being there but as the technology improves they will be just as effective. Just saving on time, cost, and effort of travel is a great advantage in and of itself.
This is one of the reasons why video games are so popular and so addictive. A subjective sense of “being there” and interacting with the others triggers our minds in the ways that are stronger and more intuitive than any “flat” experience. The entertainment industry and the media have grasped this and are embracing VR as the new platform. It represents a qualitative jump comparable to the difference between a typed line entry computer and a smart phone.
That is the vision behind Virtualitics. It is at the intersection of big data and virtual reality. It combines modern data analytics and machine learning with the data visualization of an unprecedented power and effectiveness. It is a platform where humans can interact with their data and their colleagues – and it may be the most natural platform where humans can interact with machine intelligence.
Welcome to the workspace of the future. It is a whole new world!
George Djorgovski is Co-Founder and Chief Virtual Reality Officer at Virtualitics. He is also Professor and Executive Officer (Dept. Chair) for Astronomy at Caltech, as well as the founding Director of the Center for Data-Driven Discovery.