Data Analysis: Balancing Utility and Security
Booz Allen Hamilton’s Tactical Athlete Leadership Board met to conclude 2020 to discuss best practices and challenges in collecting data in training programs. With the understanding that it is vital for the military to continue advancing human performance and training protocols, board members from across the industry, who are experts in sensor technology, coaching, and data science, came together because of the benefits of collecting data from today’s warfighter can:
Sensors that collect data off the bodies of warfighters continue to improve. We know this because we’ve seen the form factor, the accuracy, and the cost of these devices advance dramatically in recent years. This means more data.
More data on the quality of sleep, nutrition, muscle response, cognitive processing, form, position, location and speed. It’s all coming. For many leaders, this data holds the answers for modifying, or improving longstanding training protocols. For others, it brings more noise. The truth lies somewhere in the middle, but there are certainly best practices to get more from the data that has been collected.
To begin with, those in the military looking to optimize the performance of the tactical athlete have to confront some important near-term challenges:
They must decide what to keep, anonymize, and organize for analysis.
They must decide where to put all this data and who has access to it.
They must commit to analyzing similar metrics over time.
This is all easier said than done. To begin with, few in the field of human performance can agree on which metrics tie to which outcomes. Academic and medical papers debate the value of heart rate variability, the definition of load, and the correlation between data, injury, and recovery. To avoid throwing out valuable data, the approach to data has been to keep it all. This immediately creates security, cost, and integration challenges.
Newer data management products that are designed to handle and organize disparate data sources have emerged to help turn various types of data into useful dashboards for leaders to review and identify patterns. These products have emerged from the evolution of electronic medical records systems, sophisticated athlete management software used in professional sports, and business intelligence visualization tools. As “data-agnositic” hubs, they serve to take various inputs and transform them into normalized data that can be visualized consistently in a dashboard designed for non-data analysts (non-technical field coaches and leadership).
In the military, another layer of security is required to keep warfighter data secure. That’s why many in the industry are eager to see the investment in what’s called “confidential computing”. The Tactical Athlete Leadership Board knows, “Security is a top concern”.
The rise of confidential computing could help. The new paradigm involves the combination of CPU-based hardware and cloud/software tools that enable organizations to create trusted execution environments (TEE) -- also called “enclaves”. These systems offer encryption of the data in use -- that’s a key point. While the data is being processed, these enclaves render sensitive information invisible to host OSs and cloud providers. Large software and hardware companies should take part in this initiative and aim to solve security issues that have long prevented different areas of the military from analyzing the same data set or developing new algorithms by leveraging another unit’s or department’s training set of data.
These advancements in hardware and software beg the question of human talent. The military, like professional sports, needs to work with human performance data consistently to start gaining the domain expertise to identify patterns and scrub out useless data. This starts by recruiting and developing data engineers to manipulate the data. Data scientists who can develop algorithms and statistical models for analyzing data over time, and data analysts who are expert in visualizing data and communicating trends.
We can be confident human performance data is coming. It’s getting more accurate and reliable. The Tactical Athlete Leadership Board recommends developing the talent to analyze it now, so the military is in the best position to optimize it in the future.