In his trademark black T-shirt, blue jeans and Converse All-Stars, Damien Patton walks up to a television screen mounted to the wall near the front door of his startup’s office.
“I don’t think this thing’s on,” he says.
He jiggles a computer mouse attached to the screen and up pops a rotating, glowing globe. Red, yellow and green lights project out of spots on its surface.
His Las Vegas startup is called Banjo. He calls this part of it his crystal ball.
“This is the world right now,” he says. “These are the things that are literally happening and we’re picking them up.”
Scouring the Internet for posts generated on social networking sites like Facebook, Instagram and Twitter, Banjo detects 50,000 unique events every day, everything from sporting events to riots to elections.
“I can tell you everything in the world that’s happening,” Patton, 43, says in his office later. If someone hits a fire hydrant and it closes a road, the only person who knows about it before Banjo is the person who was driving the car and posted a quick photo, Patton says.
Banjo’s Las Vegas office is inside the InNEVation Center in the southwest Las Vegas Valley.
Founded in 2011, its engineers use a complex, self-learning algorithm to collect photographs shared publicly on social networks like Twitter, Facebook and Instagram. The mission of Banjo’s engineers is to create something of a worldwide heatmap, detecting areas buzzing with an abnormal increase in social network activity.
“And in real time,” Patton said. “It’s so much information, it’s painful.”
The application began as a consumer product aimed at helping users keep track of nearby friends, but after its user base surpassed 3 million in 2012, Banjo expanded its aim to include brands, like the BBC, Fox and NASCAR.
A Desert Storm veteran who ran away from home in California at 15 to travel the country, Patton built Banjo in 2011, initially setting up shop in Redwood City, Calif. His jaunt into the world of software followed careers as a volunteer crime scene investigator and seven years on a NASCAR pit crew.
Banjo found its funding — more than $20 million of it — from various investor groups, including the VegasTech fund.
Twenty-five of Banjo’s employees came to Las Vegas in 2013. About 30 engineers and data scientists still work in Silicon Valley. But it’s been easy to stay in touch.
The Banjo office is equipped with giant televisions loaded with cameras and live-streaming software. During a tour of the system last week, Patton walked up to one displaying a live stream of the Silicon Valley office, where rows of employees sit in front of computers.
He clicks “unmute.”
His employees turn their heads, hearing the booming voice of the boss.
“See? Everyone turns around,” he says, laughing at the science fiction of it all. “Nah, it’s a test! It is just a test. Settle down. Go back to work!”
The setup may seem a bit “Big Brother,” but Patton says the “virtual windows” have streamlined communication.
“When I’m white-boarding, they zoom in on my drawing,” Patton says, pointing to the board. “The collaboration is like you’re in the same room. There’s a lot of monitors in there.”
Patton has a name for finding hot spots on his crystal ball globe: world detection.
Sitting under dim blue lights and across from a bank of flat-screen television monitors tuned to news networks all over the world, engineers use Banjo’s complex algorithm to sift through millions of posts on social media, searching for relevant information.
With a Banjo account, clients can view a stream of aggregated social media content pegged to a specific geographic region or venue. That means a Banjo user can get all the news from the Super Bowl or Syria with a few taps on his iPhone.
Lately, the company has been focusing entirely on its work with brands, like Anheuser-Busch and NASCAR. Using photo recognition technology, its systems can, in a split second’s time, sift through millions of user-generated photos, log when and where they were taken — and tell Banjo what was in the photo.
That means Banjo can tell you if someone is drinking a can of beer made by Anheuser-Busch in a photograph without any mention of the company.
It works like this: The Banjo system uploads photos of company logos and chops them up into thousands of tiny triangular bits. As the system moves through content, it’s studying and defining what’s actually inside the photo.
That means a Banjo staffer can zoom into a specific location and ask the application to bring up every photo of Anheuser-Busch products taken in the last five minutes. The technology allows companies to find out how the density of their product compares to competitors’ in certain markets, allowing them to make better advertising decisions.
“Every single business in the world needs a crystal ball,” said Patton.
This isn’t the first time a tech company has used complex software to analyze the content inside a piece of media. YouTube uses a system called Content ID to scan videos for copyright infringement, comparing new uploads against a huge database of songs and images submitted by copyright owners.
Banjo wants to do the same thing to detect breaking news events, like fires, floods and riots.
As for the future, Patton can’t say where Banjo is heading. It may take on many different versions, like a modern version of the emergency broadcast system. Or maybe the next evolution of cable television. Maybe something more sci-fi.
“Nothing beats real time, except the future,” Patton said. “But if you know your history, and you understand the present, you can predict, within reason. That’s what we’re doing.”