Priming PEO EIS for data science

The Army is putting a priority on data, and PEO EIS is doing its part by educating its workforce about data science.

At a Jan. 28 Lunch & Learn hosted by Applied Cyber Technologies (ACT) — part of PEO EIS’s Defensive Cyber Operations portfolio — ACT data scientist Mark Orwat, who has a doctorate in computer science, presented “Data Science 101” for members of the workforce. Data has always been abundantly available across the Army, noted Orwat, but the service has only recently begun prioritizing data analysis, with PEO EIS program offices like ACT and Army Data and Analytics Platforms (ARDAP) leading the way.

The following are eight takeaways from Orwat’s presentation:

1. Data science is about telling a story. The term may conjure up images of technical experts manipulating data in complex ways, but data science is even more about enabling people to tell stories that make numbers meaningful.

2. Anyone can be a “citizen data scientist.” While traditional career roles in the data world have included engineers, analysts and scientists, a new category has emerged enabling anyone of any rank, as well as those with limited data knowledge, to develop data sets and use tools to tell a story. “We’ve all done data science on some level,” said Orwat, citing statistics class as an example. “You’ll be surprised at just how much you know.”

3. Ensuring clean, accurate data is important. As data streams are brought together, quality can suffer due to duplicate records, records that contain conflicting data, records with incomplete data, or records with data considered invalid because it does not conform to a standard. Everyone has a responsibility to ensure data and metadata entered into systems are as complete and error-free as possible.

4. Besides being of “decision-quality,” data needs to be conditioned to determine what is useful and what is not. Currently, people are putting huge amounts of data into a single common data repository, but if unproductive data continues to be added to the platform, the system will get overwhelmed.

5. Big data is more than about just volume or size. Big data is characterized by the five V’s: volume, variety, variability, veracity and velocity. “PEO EIS has a huge interest in big data,” said Orwat.

6. Gaining insights from data is good, but using it to make predictions is the wave of the future. Right now, PEO EIS is successfully analyzing data to create reports and dashboards to gain business insights. As the organization gains more experience and develops more toolsets, it will be able to assess potential future scenarios and use artificial intelligence to predict behavior in unprecedented ways.

7. PEO EIS aims to put the Army Data Plan into action in four phases. Phase one (“Set the Conditions”) will begin the second quarter of fiscal year 2021, with the program executive officer issuing an operations plan (OPLAN) to program managers. As part of the OPLAN, PEO EIS leaders will initially assess the current environment and establish a framework for sharing understandable and usable data assets. Second, leaders will adapt and build the framework under ARDAP’s lead. Third, they will implement the framework across PEO EIS, facilitating adoption by all program offices. Lastly, leadership will establish systems to maintain the environment.

8. ACT continues to mature its data capabilities to better enable Army Cyber Command’s cyber protection teams (CPT). Leveraging data and predictive analytics to detect adversaries using precise, pre-established priority intelligence requirements will better posture the CPTs to root out future attacks.