Posted by
You’ve probably heard a bit about big data in recent months. Chatter abounds about the enticing possibilities such prodigious data collections offer. But what, really, is in store for owners and users of big data sets?
Spotlight: On-demand video
AI Explainer: Foundation models and the next era of AI
Explore how the transformer architecture, larger models and more data, and in-context learning have helped advance AI from perception to creation.
Curtis Wong (opens in new tab) knows.
He should. Wong was the Microsoft Research scientist who gave the world the WorldWide Telescope (opens in new tab), used by legions of astronomy fans fascinated by the informative, fun experience offered by a virtual telescope that delivers seamless, guided explorations of the universe.
On April 17, during Microsoft Research’s Silicon Valley TechFair (opens in new tab), he is demonstrating a project called Holograph, an interactive, 3-D data-visualization research platform that can render static and dynamic data above or below the plane of a display, using a variety of 3-D stereographic techniques.
“Holograph is yet another step in our effort to democratize access to large data sets,” Wong explains. “It started with WorldWide Telescope, in terms of making all that data available to everybody.”
Holograph brings the data visualizations of the heavens down to terra firma.
“There’s quite a bit of spatial and temporal data here on Earth,” he says, “so that led me to work with the Office (opens in new tab) team to develop spatiotemporal data visualization in Excel (opens in new tab). That enables anybody with any kind of data, like location or time, to easily plot that data onto a map or globe. If you select ‘date,’ and you have date or time information, you’ll be able to watch that data play out over time.”
Holograph enables users to examine various types of dynamic data sets—easily and naturally.
“If you have 3-D data, you could move your head around and look at things as if they were physically there,” Wong suggests. “That’s a natural way of observing something. If you want to reach in and select something that’s interesting, you could find out more information. Sure, you could do that on a flat screen, but on really complex, non-two-dimensional data, it becomes more of a challenge.
“What we’re doing with our Perceptive Pixel display and things like Kinect (opens in new tab) is tracking where your head is, where your point of view is, and dynamically altering that data. The touch capabilities of the Perceptive Pixel display allow us to understand where you’re touching and then have that touch point connect with a piece of 3-D data represented below the plane of that screen. We’re rendering 3-D data both statically and dynamically in what looks like a virtual box below the glass, but you can see the dimensionality of the data.”
If this all sounds a bit sci-fi, well, it should. Wong’s been learning to think along such lines for quite a while now—with the help of some genuine inspiration.
“I decided to start exploring this from working with Jim Gray (opens in new tab) about 15 years ago,” Wong says. “It became obvious from hanging around Jim that the world was going to be inundated by really, really large data sets. I wanted to make that accessible to everybody, so everything I’ve been doing is to try to be pre-emptive about that.”
And, it turns out, the combination of the spatial with the temporal unlocks entirely new possibilities.
“We have so many mobile devices,” Wong notes. “We have sensors everywhere that are capturing so much information. How can we make sense of that? And not just get a picture of it, but make sense of it over time? Hence, I’ve been advocating that to be able to do temporal analysis is really important. That’s why I’m doing this.”