Relational graph databases can help staff, structure and deliver projects on-time and on-budget, writes David Meza
David Meza is a Sr. Data Scientist at NASA headquarters. With a background in computer and information sciences, he has held various roles at NASA for more than 20 years, from IT management and Chief Knowledge Architect to workspace engineer, overseeing 12,000 computer systems at the Johnson Space Centre.
In December 2017, President Donald Trump signed Space Policy Directive 1, authorizing the United States’ next lunar visit. Here at the National Aeronautics and Space Administration (NASA) we got to work on the Artemis program, a $30 Billion plan to put the first woman and the next man on the moon by 2024.
As we started preparing to go back to the moon — and on to Mars — we realized that we need people with skills that are unique to our projects and not found in abundance. Beyond that, we identified a need to realign skills to better support our mission needs. The solution: A graph database.
The beauty of a graph database over a traditional database is that it is uniquely built to analyze relationships, and the model (or schema) is not fixed. This makes it abundantly useful for NASA, and for large capital project organizations, too. In the article that follows, I’ll introduce the concept, explain some use cases, and walk through the early implementation phases.
What is a graph database?
A graph database is purpose-built to store and navigate relationships. It uses nodes to store data, and edges to store relationships between that data. If you played with Tinker Toys as a kid, this will be easy to visualize. The node is like a wood wheel, and it contains data, like a person’s name, occupation, or location. The edge is like a stick, and it contains a relationship, as well as information about that relationship. There is no limit to the number of relationships a single node can have, making graph databases scalable, powerful, and very useful. Unlike traditional databases, a graph database is literally built to analyze relationships.
Unparalleled visibility, uncommon insight
Using a graph database, leaders can look at a capital project from a truly relational standpoint. Consider a database in which the node for each employee is connected to nodes detailing their education, training, projects and work roles. The node for each work role is connected to nodes detailing the required personal characteristics, skills, abilities. The node for each project is connected to nodes detailing cost, schedule and operability outcomes. Suddenly, it’s possible to analyze which people, training, and personal characteristics correlate with your most predictable projects. That’s the power of a graph database.
Four Use Cases for Capital Projects
Staffing and Skill Sourcing
A leadership team is staffing a one-of-a-kind project with an important, unprecedented role that requires an uncommon skill set. There are no obvious internal candidates for the position. Using a graph database, the leaders can input a list of skills required for the role, and find employees that match (or very nearly match) the requirements.
A global organization is working to assess the strengths and weaknesses in its existing project delivery systems. Why does one project fail, and another succeed? Using a purpose-built graph database, the organization can identify which methods, tools and people are engaged on successful projects, and which are engaged on those that fail.
Human Resource Development
Leaders in the human resources department of a struggling multinational company are working to identify which in-house training programs to develop. A graph database can deliver a data-driven skill gap analysis along with insights into training paths, career plans and adjacent skills.
An ambitious young company is looking to adopt Advanced Work Packaging, but leaders do not have a reliable, objective measure of the organization’s maturity. Using a purpose-built graph database, the company can identify the markers of maturity — skills, tools, metrics, etc. — and then benchmark the company against those markers. Once the company has reached the baseline markers, it can be considered mature enough to begin the AWP implementation process.
Getting Started with Graph Datasets and AI
The move to graph databases must be driven by leadership executives and led by the organization’s information technology professionals. Data about employees will have to be derived from internal records, and data stewards will need to wrestle with ethical and legal questions in addition to the primary technical challenge.
A skilled data engineer will have to organize how the data will fit into the model, and the company will need teams to both build and analyze the databases. Finally, many organizations will want to build an interface to get the most out of their data. It’s a big project, and one most likely to be used by large organizations with substantial resources.
Designing, building and leveraging a graph database is a big challenge for bold, well-resourced capital project companies that want to position themselves for sustainability and growth in a rapidly changing economy. Doing so will yield a strong competitive advantage in the years and decades ahead.