For the Air Force, big data analytics in ISR is a team sport

The Air Force is struggling to meet unexpected demands for greater intelligence, surveillance and reconnaissance collection—not to mention analysis of the data—amid tighter budgets.

Two distinct issues concerning the service in terms of ISR are sequestration, which constrains resources, and increased operations, according to Jon Kimminau, analysis mission technical advisor for the Air Force’s Deputy Chief of Staff for ISR.

During an event hosted by AFCEA on Wednesday, Kimminau noted that, around 2010, the military thought it would be getting out of counterterrorism operations in permissive environments. But with the rise of ISIS, that hasn’t happened. Counterterrorism operations have significantly increased, and plans now call for them to increase even more in the future. The military has announced that it will be increasing its ISR flights—referred to as combat air patrols, or CAPs—by nearly50 percent by 2019.  

The other challenge is making the most effective use of the ISR data collected. “For ISR, we essentially get data, turn it into information, transform it more into knowledge, and make it something that decisions can be acted upon,” to give an advantage to decision makers, Kimminau said. But the focus has been heavily on the data-to-information side and “really sloping down towards the knowledge and decision quality advantage,” he said. “That’s not where we want to be.” He added that “we want to shift our workforce and resources and systems from being heavy on the data information side to being heavy on the production of knowledge side.”

The answer to many of these problems is greater use of data analytics. One vision is to take all the data associated with various intelligences, such as signals and geo-spatial intelligence, digitize it and put it in a cloud architecture, he said. This will allow for better data analytics and thus better and more informed decisions for warfighters. It will also allow collection to be driven by analysis as opposed to collection being driven by requirements.

“We have to revolutionize how we do analysis because that’s the engine of intelligence,” he said, adding that this will allow the force to accomplish its goals with current resources and personnel and a cloud architecture and data analytics will be a huge part for where the force wants to go in the future of ISR.

Kimminau also dove into three “nagging” questions surrounding the adoption of big data analytics and where the force wants to go in the future: what the Air Force and its personnel knows about big data analytics, how does it get to big data analytics from where it is currently and, lastly, whether the shift to big data analytics can evolutionary or must it be a transformational change involving aspects that are not known or officials might not have conceived of yet.

One approach is greater collaboration. He used the example of Amazon—not in the sense of thinking of a particular item and searching for it, but rather  the function that tells the shopper, or in this case, an analyst, what others who searched for that item were searching for and how it was rated by others. This analytical capability, powered by big data analytics, would greatly improve the analysis environment, he said.

Getting to that point is something the Air Force needs to accelerate, he said. Presently it takes about five years to deploy new systems, which is not where the force needs to be. Rather, he wants to get to months as opposed to years to which the force is making progress. “We’re fighting a war, we’re at max capacity and tempo and we can’t let up off of that while we are changing to this new environment,” he said, comparing the challenge to building an airplane in mid-flight. The trick will be figuring out how to do both—building as they go while understanding the process. 

Big data also changes the old, “classic” analytical framework, which was like finding the pieces that fit into a puzzle. With big data, all the data is there, so an analyst might start with a question rather than a requirement. They might look at a particular area for strange behavior, anomalies, etc. and then deliver the information, Kimminau said. Big data provides a revolutionary approach toward discovery in the intelligence field—not just the discovery of data, but discovering something that analysts didn’t even know was there.

Finally, he noted that big data analytics is going to be a team sport. It’s not just a job for intelligence analysts. It involves an analyst, a data scientist that understands how to work with data and a knowledge manager that understands how to package information and cloud architectures.