Companies offer life preservers for analysts drowning in data
TAMPA, Florida – Every year, U.S. Central Command collects the video equivalent of 325,000 feature length movies and enough signals intelligence to replace 5.5 million songs on an iPhone.
“It presents a massive challenge for us,” Lt. Gen. Charles Brown Jr., Central Command deputy commander, said April 25 at the 2018 GEOINT Symposium here. “Who has time to sift through all that to find the golden nuggets of information?”
Because this data problem is widely recognized, companies filled the Tampa Convention Center April 22-25 with tools designed to ease the workload.
Harris Corp. demonstrated Distributed, All-source Geospatial analytics Resource, or DAGR, a browser-based application that uses the company’s Hydra enterprise network architecture to help government analysts comb through data from multiple sources including space-based, airborne and ground sensors.
DAGR identifies normal patterns of behavior and alerts analysts to activity they might want to home in on, like three or more vehicles meeting in one spot or traffic between two known “bad guy locations,” said Josh Nauman, Harris chief solutions engineer. “Our software recommends analytics based on all the data that’s out there similar to the way Amazon and Netflix recommend products. We are trying to make it easy for analysts to grab those complementary data sources.”
Across the exhibit hall, DataRobot, was showing how someone could apply its automated machine learning platform to satellite imagery. An intelligence analyst could information on ISIS facilities in Raqqa, Syria, for example, to predict which buildings in a neighboring town were likely to be key ISIS infrastructure.
“You put in your data and tell it what you want to predict, like ISIS infrastructure, said Kevin Stofan, DataRobot customer facing data scientist. “It looks at variables and data, runs open source models, picks the top models and makes predictions.”
Since it was founded in 2012, DataRobot has focused largely on the commercial market, helping banks, insurance and health care companies build predictive models. In 2016, the company began moving into the government market. It now has contracts with the Defense Department, Department of Homeland Security and intelligence agencies in addition to the backing of In-Q-Tel, the venture capital fund that invests in technology with intelligence applications.
Not far from DataRobot’s booth, Mapbox was advertising its platform, which helps developers bring basemaps, from satellite imagery to street maps, into “anything and everything,” said John Dombzalski, Mapbox global government strategy director.
Mapbox commercial customers include Snapchat, the Weather Channel and CNN. Lately the company’s U.S. government business has been expanding as customers including the Air Force, National Oceanic and Atmospheric Administration, and Centers for Disease Control and Prevention use Mapbox to bring location data into various applications.
At GEOINT, Mapbox was showing potential government customers how they could use its Atlas Server to visualize dense stacks of data and swipe between imagery layers. Mapbox works with multiple cloud computing services including the intelligence community’s Commercial Cloud Services, Dombzalski said.