Building a Network View of Data
Linked Data versus Big Data
I’m a fan of Michael Arena’s work, along with others doing work in organizational network analysis (ONA). If you don’t follow Michael you should. Besides being a super smart guy, he is a kind, approachable, wonderful human being. One of the most powerful things I’ve learned from him is the distinction between human capital and social capital.
A human capital lens focuses on individual people, their skills, their productivity, and the value they bring to an organization. There is nothing wrong about focusing on your organization's biggest asset but it can’t be done in isolation. We are not just people, we are people with relationships forged over time through diverse experiences. We don’t work alone. Rather we are a part of teams working together towards common objectives. A social capital view considers how individuals work and function as a team.
Individuals linked together can create more value than the sum of their parts. When our teams function well, we gain energy and ideas from each other and find motivation to tackle tough problems together. When they don’t function well, the opposite happens and we lose that motivation and energy.
The idea of people linked together in many ways is powerful. I’m a big believer in the power of networks of people who can contribute much more than the individual contributors. ONA is a tool to help understand the networks in organizations and how to improve them to increase social capital. I’ve successfully used ONA in various situations but that may be a story for another time.
I’d like to take this analogy of a network of people and apply it to data sets. Just as individuals become more valuable when they are linked to other individuals, the value of a set of data becomes much greater when it is linked to other sets of data. Linked or networked data can provide additional insights and information that you can’t get from individual datasets.
From a data perspective, the challenge is not about getting the most data that you can get. It is not about getting “Big Data” which doesn’t seem to have the buzz now that it once had, perhaps swallowed up in the AI buzz. I’ll take linked data over Big Data any day. Too often, people data resides in different systems, in different formats that make them challenging to link. Linked datasets allow you to move up on the analytics maturity curve (such as this example shared by Keith McNulty) I would venture to say that linked datasets are a requirement for advanced analytics.
Take the example of an employee. While you may have demographic and job information in an HRIS for all employees, you don’t have everything. Their training data might be over in the LMS system. Their applications, interviews, and resumes are over in a talent acquisition system. Their performance reviews, development goals, compensation history, and skills could be in still other databases or systems. If you want to get a holistic view of an employee, you have to be able to link these disparate sources of information together.
So how do you get to a network view of datasets? Two critical elements are data engineering and data governance.
That means people analytics teams need data engineers. Data engineers are crucial in being able to link and connect different sources of data. I would invest in good data engineering talent before investing in good data science talent. Of course I’d love both but if I could only hire one, I’d start with the data engineering. There is a difference between the data engineering of ETL processes to feed production-level dashboards and the data engineering to bring together datasets to explore potential hypotheses in advanced analytics applications. Both are important but no matter where you are on the analytics maturity curve, you will need data engineering skills on your team.
In addition to data engineering, data governance is key. Whether it is driven by the people analytics team or not, data governance efforts need people analytics as a crucial partner. If you don’t have consistent definitions and data standards, you can’t link your data sources together and lose the potential value you can get from linked data.
Just as a social capital view of people and organizations leads to valuable insights, so too does a network view of your data lead to more insights than you can get from disparate, disconnected sets of data. Linking your data together should be a top priority for any people analytics team.
If you want to be notified of new articles, you can subscribe for free to this Substack. I aim to post every few weeks and welcome thoughts, questions, and comments.