Europe’s Largest Data Observatory Turns Big Data into Big Images

Because humans are good at recognising patterns in images.

Nov 3 2015, 10:00am

David Birch, the architect of the Data Observatory, at the center of his creation. Image: Emiko Jozuka

Humans are pretty visual creatures. Since caveman times, we've scrawled stone walls with pictographs and paintings both to reflect and to understand our changing times. Fast-forward several centuries, and we're doing something similar with Big Data.

Opening Tuesday at Imperial College London's Data Science Institute, the KPMG Data Observatory is Europe's largest ever data observatory. It's a circular wall of flickering screens that transforms data into images and allows users to visualize everything from global migration patterns and cryptocurrency transactions to commuter movements all the way over in Shanghai. It's intended for use by big business, the medical sector, and students.

"This data observatory is a place where you can see patterns in the data," Mark Thomas Kennedy, director of Imperial Business Analytics, told me as we stood at the opening of the cavernous observatory.

The dots reveal the exits for metros across Shanghai in China. Image: Emiko Jozuka

The igloo-like structure took around a year to build and three months to assemble, according to David Birch, a researcher at Imperial College London's Department of Computing.

"It's a bit of a beast to maintain," he told me, as we stepped inside. Birch swiped his tablet, making multiple screens light up with everything from a sprawling Bing map across 130 million pixels to visualizations tracing how bitcoins are traded. "It's an immersive environment so it can be a bit dizzying," added Birch as the screens morphed into different data skins. The observatory has 64 monitors powered by 32 computers, and is equipped with 313 degrees of surround sound.

Inside the Data Observatory, it feels like you're at the center of a contemporary art exhibit. But instead of trying to work out a narrative for a conceptual piece, you're doing is parsing meaning from immense amounts of data points.

Kennedy argues that the idea of the observatory is to "take advantage of years and years of evolution," and exploit the fact that "humans are really good at recognizing patterns visually."

Looking at the evolution of the bitcoin blockchain. Image: Emiko Jozuka

Made in collaboration with researchers from Imperial College London and investment firm KPMG, the observatory is meant to be a bit like a predictive crystal ball for big business. They could hand over their datasets, then let the observatory visualize everything so that execs can spot patterns and trends emerge. This lets them explore what could happen before they actually throw their business to the market wolves.

"The more we can turn data into something visual, the more people can actually respond to what's going on in the world, and take decisions that move organizations forward," said Kennedy, suggesting that businesses would be more willing to take risks when they could see patterns emerge from the decisions they had made previously, or were about to make.

Using the Data Observatory to spot patterns between patient data. Image: Emiko Jozuka

Mazhar Hussain, director of digital & analytics at KPMG, proposed how banks could plot their branches in any given country on their data visualization, and from there receive information from each of the locations, and make better decisions on where to open new branches, or close existing ones.

Researchers predict that it'll be another couple of years before the tech behind the Data Observatory spreads beyond the Imperial lab. In the future, however, they want it to operate in real-time, and have people—students, business execs, and medical staff— interacting with similar observatories on a regular basis within their own institutions and companies.

"Looking at these visualizations is a way to get signal from noise. [...] Data transactions are happening all the time, but there's not visibility," said Kennedy, citing the example of money laundering services and scam attacks. "The ability to track all this visually is a step towards building tools that track something when you're not looking at it."