NASA is implementing a graph database to serve as the foundation of its people analytics. It will give managers and employees insights into skills and training needs.
NASA is taking a more centralized management approach as it prepares to take on ambitious missions, such as sending astronauts to Mars or back to the moon. It is a project with implications for HR and how it conducts people analytics.
The space agency wants a better view of its workforce and current skills as well as what skills it will need in the future.
To do this, NASA's HR department has started using a graph database platform from Neo4j Inc. Graph databases map relationships between data and can uncover patterns that are hard to find using traditional people analytics techniques, which often rely on relational databases. David Meza, a senior data scientist for the chief human capital officer at NASA, is in charge of the project.
Prior to his role in HR, Meza was the chief knowledge architect at the Johnson Space Center in Houston and has nearly a decade of experience with graph technology. He had a good idea about what it could do for the space agency's HR department.
When considering people's relationships, Meza said, "What more type of graph problem out there is there than people?"
But the question for HR, broadly, is whether graph database technology will be mainstream and widely adopted by vendors that sell people analytics platforms.
Josh Bersin, an industry analyst and head of Josh Bersin Academy, said graph database technology "is starting to emerge as a tremendous technology for HR." But HR departments want off-the-shelf products, "so it's coming slowly," he said.
What more type of graph problem out there is there than people?
David MezaSenior data scientist, NASA
Bersin noted that some HR vendors use graph databases, including ADP's NextGen HCM and Workday's skills cloud.
"I think in 10 years, all HR systems will use some form of graph technology for sure -- but it's slow to build because it takes vendors three to five years to build off-the-shelf products that capture the hearts of conservative companies," Bersin said.
Holger Mueller, an analyst at Constellation Research, is more skeptical of the future of graph databases in HR. He sees the potential graph technologies could have with organizational charts and other hierarchies. Still, a complex task like mapping skills across the workforce is "critical and hard, [and] a graph database doesn't make that easier," he said.
HR people analytics
Graph databases capture and store relationships that show interdependencies, hierarchies or nodes and how they link to one another, not as data in relational database rows and tables. Relational databases don't surface relationships between many things or entities, "so when you try to model a complex organization, it's really hard to do," said Amy Hodler, director of graph analytics and AI programs at Neo4j.
Currently, NASA uses relational databases for its HR data. In many cases, NASA managers keep spreadsheets about their employees and include information about them, Meza said.
Meza wants to tackle questions such as: "Where are our skills? How is our workforce aligned to our work roles?" The goal is to use graph database technology to modernize its people analytics.
A summary of the key differences between graph and relational databases.
Meza started by breaking the problem down into knowledge, skills and tasks at their most basic level -- what makes up an occupation and how that relates to employees with certain knowledge and skills. But that can also be extrapolated to training and then to a project or program, he said. From these relationships, "you can really start to glean some information from your employees."
The graph database would be used both by employees and managers. For employees, a graph database could help them find occupations, work roles and assignments at NASA "that are in my interest, from an employee standpoint," Meza said. Employees can also discover what training they need for those roles.
From a manager's perspective, the graph database will help find the people they need for a particular task. It will also help managers discover "what skills am I losing … do I see a pattern happening here? Am I losing too many of my data scientists or too many of my robotic engineers? And why am I losing them?"
Meza has been at work on this project since late 2019. His ultimate goal is to create a central hub of data "that would allow individuals to query for whatever type of question they're looking for."
Patrick Thibodeau covers HCM and ERP technologies. He's worked for more than two decades as an enterprise IT reporter.
