Defining Data Worlds

Data worlds are virtual immersive spaces generated by data representing a complex system.

The approach is to bring all the necessary data, into one context, to understand and manage that system, whether big or small. This contrasts from the usual data visualization analyses that focuses on one aspect of the system at a time. While this often stimulates a deeper insight into that aspect, understanding and management require the greater context to balance the priorities across the whole. The two approaches – breadth and depth – are complementary.

Managing a Large Manufacturing Operation

Consider the following example. In 2010, I conducted a study of Western Digital Corporation, examining their use of analytics within their manufacturing processes. The system was multiple complex assembly lines that produce thousands of disk drives per day. Their engineers were collecting every tiny piece of data and integrating it into their data warehouse. Hundreds of engineers were obsession with quality control of that system, thus constantly performing analytics. The motivation was that, arising from countless root causes, a defect could be propagated into thousands of drives within hours. A cheap component could cause costly rework of faulty drive prior to shipment and much greater losses if shipped to customers. What if… all those engineers and their managers could collaborate within a data world that integrates the data and its analyses into one context. Not replace their analytic infrastructure, but consolidate it.

Thinking like a Themed Park

The Immersive Worlds Handbook by Scott Lukas, an architect/designer, is a fascinating book focused on physical places (not virtual spaces) where people can be get wrapped up (i.e., have an immersive experience). Examples are amusement parks, restaurants, casinos, resorts, shopping malls, and even cruise ships. He defines a world as a complete, diverse and consistent place, with a history and culture, inhabited by persons. It is ever-changing, evolving by relationships and forms of interconnections.

It is surprising how many topics and issues from Lukas are relevant to data worlds. Hence, it is useful to think of a data world as a well-designed themed park, much like Disney World. Instead of providing entertainment, the data world lays the elements of the system over a vast plain with data spouting and refining at appropriate locations. The data is refined via analytics into information-rich data objects, which generalize essential relationships. Unlike being at Disney World, we can ascend vertically to view the system at any level of aggregation.

Similar to the Enterprise Data Warehouse

The spirit of designing a data world is similar to the enterprise data warehouse, whose focus is the entire organization across functional areas, as a single integrated system. Data from one functional area (subsystem) is interrelated to that of another. It is those ‘cracks’ between functional areas where the greatest opportunity lies. Over the past two decades, data warehouses in all the major corporations globally has proven this premise. Essentially, the data world approach would leverage the data warehouse to increase its ‘active’ role in managing the corporation.

Personify the Data

If you were a megabyte of your data, how would you like to look and behave? Put on a yellow dress or tan slacks? A Congo Line with your friends? With whom would you like to socialize? Data personification within a data world may be an effective way of introducing complex analytics (clustering, regression, PCA/SVM), as the personalities (appearance and behavior) for the data.

Sensing, Not Just Seeing

Within data worlds, we should use the term sensing rather than visualizing for two reasons. First, a lesson from video gaming is that sound is essential to the immersive experience. All activity within a data world should have an appropriate sound behavior. Not only does it complement the visual experience, but it also adds an ambient sense of what is happening in the local environment that you can not see. As VR technology matures, we need to think…

Sensing = Seeing+Hearing+Touching+Smelling+Tasting+…

Second, an immersive experience can be deeper than just compare and contrast as with 2D visual analysis. You can sense, as almost a subconscious feeling, what is in harmony and what is in discord. Let the Force be with you! Seriously, we may be able to push the immersive experience to deeper (even scary) levels of understanding complex systems.

Other sections deals with Designing Data Worlds and Building Data Worlds.