Drones are really data vacuums. Here’s how one startup plans to clean up


Drones may still be a bit of a novelty in the skies, but cottage industries are already springing up around expectations that their impact will be enormous.

Fleets of drones are already being used in fields such as law enforcement, agriculture, real estate management and construction. Many other applications are expected to burst forth as standards and regulations take shape.

Companies such as Flyspan Systems Inc. are getting in on the ground floor. The Los Angeles-based startup is developing analytics software that helps owners of drones fleets deploy their airborne devices more efficiently. Now it’s adding image recognition to alert drone owners to problems the machines spot from the sky. Up next: artificial intelligence-driven video recognition.

Brock Christoval, FlyspanFlyspan launched in February 2014, just two months after Amazon.com made headlines by announcing an audacious plan to deliver shipments directly to customers via drones. “That’s when the industry really took off,” said Flyspan Chief Executive Brock Christoval (right).

Christoval and co-founder Vinny Capobianco immediately saw the potential for commercial applications. Drones can easily reach some of the most inhospitable places on the planet, which makes them ideal for tasks like diagnosing problems on oil rigs or gas pipelines and checking cell phone towers.

Data from above

Drones are data capture machines. They can stream sensor data and video feeds to central monitoring facilities for instant analysis. That’s where Flyspan fits in. It’s building a platform to capture live data from drone fleets and massage it into information its customers can use to optimize efficiency and spot problems. Its secret sauce is software that harmonizes video and data streams to enable decision-makers to precisely pinpoint the location and time of the drone fly-over.

“We let pilots do the job of flying drones so the decision-makers can sit in an office and make informed decision,” Christoval said.

The challenge is separating the wheat from the chaff. Drones deliver a lot of data, but not all of it is important to every customer or application. Altitude, temperature, geographic coordinates, wind velocity, and battery voltage may be background noise to someone monitoring video to look for disruptions in the pipeline, but vital data to someone performing predictive maintenance.

Filtering the right data out of the stream is typically done with data integration software and involves a time-consuming extract/transfer/load process. In Flyspan’s world, there’s no time for that. The company instead chose a data preparation tool from Trifacta Inc. that’s designed for data transformation in environments that don’t use big data platforms. The company calls the preparation process “data wrangling,” and a key feature of its toolset is the ability for users to visualize data as it’s transformed and imported into an analytics engine. “It’s an intuitive product,” Christoval said. “The visualization is responsive and dynamic.”

Filtering on the fly

That fit the bill just about perfectly for Flyspan. “In a traditional data lake, the problem is visualization. It’s hard to tell in a CSV or tab-delimited view where to make data start and stop, and where to take out extraneous data,” Christoval said. “With Trifacta we’re able to visualize all that.”

Filtering data isn’t the only challenge. Drones may cross time zones in the process of making their rounds, and the data may be sent to a control center in yet a third time zone. Normalizing these factors through offset calculations is critical to knowing exactly where the drone is located as it conducts surveillance. “That can be the most gruesome thing to work with,” Christoval said. Trifacta makes this simple through the addition of a column that adjusts for time variances.

Flyspan believes it’s on the cusp of a drone-powered revolution. “The industry is about to explode,” Christoval said. His company is in late-stage development with technology that can analyze videos to look for anomalies such as water leaks in a pipeline or missing transceivers on a cell phone tower. The company is also exploring opportunities to create mobile hotspots that are essentially traveling launch pads for drone fleets.

That’s a big agenda for a company with only six people, but Christoval said the sky’s the limit. Flyspan is in the process of raising funds, and it already has paying customers. “We’re totally passionate about this,” he said. “We’re excited about being able to change the world.”

Photo by Peter Linehan via Flickr CC