Army data scientists build customized analytics to empower agile decision making

By Maureena Thompson, Army Futures CommandAugust 29, 2022

Rick Haddad, AFC G-3/5/7 Requirements Branch Chief, points to a timeline in an analysis portal.
Rick Haddad, Army Futures Command G-3/5/7 Requirements Branch Chief, points to a timeline in a decision analysis and prioritization portal. Haddad enlisted the help of Army Futures Command Data and Decision Sciences Directorate staff, including Alan Monaghan, left, and Lt. Col. Ralph Parlin, right, to develop the portal, which consolidates and refines large amounts of diverse data into easily actionable information. “To modernize, you must be able to integrate; to integrate, you’ve got to be able to communicate rapidly and effectively,” Haddad said. (Photo Credit: U.S. Army photo by Austin Thomas, Army Futures Command) VIEW ORIGINAL

AUSTIN, Texas — Members of Army Futures Command’s Data and Decision Sciences Directorate are eager to help dispel common misconceptions about data science.

“Few people understand everything needed to enable business processes with technology. Our technologists usually do not fully understand the customer, and our customers do not understand software and business process design principles,” shared Lt. Col. Ralph Parlin, the directorate’s Analytic Support Division chief.

“Therefore, building effective decision support analytics require an iterative approach.”

Rather than building tools or analytics in isolation, Parlin clarified that his teams are tightly integrated with their customers, who drive both the requirements and the priorities. Using a cloud-based analytics platform enables the teams to quickly prototype features articulated by their supported customers. All the while, they ensure each tool or pipeline feeds into a shared database that enables decision support.

“We work with the right people and are able to iterate fast enough to keep the customer highly engaged in the development process. Moreover, our team has strong enough fundamentals that each project contributes to the command’s decision support system.”

By leveraging an information-rich network of inputs and capabilities, the directorate is able to iterate toward an “optimal solution that allows senior leaders to make decisions faster and more thoroughly,” explained Senior Data Scientist Alan Monaghan, also a member of the Analytic Support Division.

According to Monaghan, a key aim of the directorate’s work is “liberating stovepipe data into more of a data environment” — a journey that frequently involves undoing “siloed, static processes” in order to collect robust and relevant data that allows stakeholders to “organize, assess and present in multiple ways.”

Parlin, Monaghan and their data science colleagues work closely with other command headquarters directorates to develop tailored, user-friendly decision support tools, which improve information storage and synthesis while also saving the Army time and money.

Monaghan notes that in the past six months, data pipelines put in place by the directorate have increased the command’s ability to aggregate, retain and incorporate data into decision making while concurrently minimizing the need to manipulate data manually.

Efforts to make data more influential and actionable do not happen in a vacuum, however, and are the product of intentional partnerships with command staff who present directorate members with compelling, senior-leader-approved problem sets and nuanced understanding of the issues and complexities at hand.

“Our directorate’s cross-functional teams consist of data scientists and knowledge management professionals supported by software developers and data engineers. We see these ‘Analytic Support Teams’ as a key element within the staffs of a ‘data-centric’ Army,” Parlin said, underscoring the importance of “rapid iteration and shoulder-to-shoulder cooperation” in developing clever applications that add substantial value.

The directorate recently collaborated with Rick Haddad, the command’s Operations, Plans and Training Directorate Requirements branch chief, to create a customized web portal that tracks and analyzes the ongoing development of requirements for priority modernization efforts.

The interactive tool, which includes a dashboard of key documents, up-to-date timelines and detailed, color-coded charts, gives multiple users access to the latest requirements data, providing a streamlined resource for information integration and decision making.

“We built a dynamic system,” Haddad said, noting that the transparency, efficiency, accountability and common operating picture offered by the portal is exceedingly useful to senior leaders, particularly when they are called on to make quick decisions or need to evaluate precisely how much progress has been made on a specific effort.

The portal is also distinctive in that it is constantly evolving and can be adjusted to incorporate leadership feedback and meet new or shifting needs. “This is invaluable,” Haddad said of the inherent flexibility.

Moving away from an outdated reliance on clunky spreadsheets, overstuffed file folders and cluttered presentation slides — Haddad hasn’t briefed from PowerPoint since the launch of the portal — has improved the command’s ability to consolidate reporting mechanisms, horizontally scale successful data models and modernize swiftly.

“We deliver support by employing software and data fundamentals that are firmly established within industry,” Parlin explained, adding that the directorate’s technical savvy combines artfully with project owner business and subject matter expertise to yield highly productive outcomes. He estimates that the directorate’s implementation of more agile and synergistic data solutions is saving the command 630 staff hours per month.

The time-saving attributes of specialized web applications mean the command can free up resources for other mission-critical activities — including ones that remain best handled by human thought processes.

“We know that if you have to spend unproportionate amounts of your time gathering the data, making sure it’s deconflicted, validating the data, organizing it into a decision support product — usually via PowerPoint — and then delivering it to a decision-maker, that type of process is slow and often inaccurate,” Parlin said. “If we can accurately capture our inputs in a consistent manner, we can automate data processing tasks. That enables our people to do what people are great at: analysis and interpretation.”

Parlin emphasized that at their core, the interfaces the directorate supports are about much more than automating number-crunching and capitalizing on new technologies; instead, they serve as conduits for “working through the information requirements of a headquarters” and enabling headquarters staff to “make better decisions faster.”

“This informs how we can execute in lean environments, if we organize correctly and we organize around this idea of being able to deliver capability through digital transformation,” he said.