Pokémon GO, Augmented Reality, and Future Number Crunching
The success of Pokémon GO shows that the public has a large appetite for digital experiences that rely on massive and sophisticated data-handling techniques. Does this mean we may eventually see more demand for systems that allow people to interact more directly with the data we now know can be delivered to mobile devices?
The success of Pokémon GO shows how the clever packaging of an engaging digital experience helps users bypass the sophistication and complexity of enabling technologies. Instead they focus directly on the gaming experience. This is partly because location enabled smartphones are ubiquitous and partly because the game-play itself is fun for so many people.
What might this tell us about how data will be accessed and manipulated in the future?
Pokémon GO relies on a variety of different data-dependent technologies which combine seamlessly to support the game experience. As the volume of accessible data around us continues to grow through expansion of real-time sensor data, instantly scaleable cloud infrastructures, and increasingly sophisticated and powerful mobile devices, will the beneficial and productive use of numeric data also become even more second nature than it already is?
A lot will depend on how we define "use." With many government open data portals, for example, data files are easily accessed via downloading. In some cases portals also provide mobile-device-friendly filtering and visualization tools. At the end of the day, though, how are the data provided by such portals actually used? Are such data files having a positive impact on decisions and actions given that users still need to jump through data manipulation hoops just to make the data understandable and actionable?
Which brings us back to Pokémon GO. Our main experience with locating and capturing these little virtual creatures is not by manipulating rows and columns of numbers but by the game software’s rendering of these numbers as simulated objects in a real background. As we look to the future of online data access, can we expect similar breakthroughs in how we interact with the vast array of data we're constantly swimming through as the “Internet of Things” continues to expand?
I'm less interested in just attaching numbers to an object being scanned via a heads-up display, like Arnold did as he walked around the bar looking for clothes at the beginning of Terminator 2. I want to reach out and "touch" the numbers and render them visually via some sort of immersive modeling tool that lets me color-code data points to represent, say, the shifting of a dependent variable overtime. I want to interact by voice and say something like, "Holding inflation constant over the next three years, show how subscription revenue grows as a function of new versus existing customers." Then I want to "walk around" this display and see immediate results of the tweaking of the underlying data.
Nothing suggested above is functionally out of reach of a sophisticated spreadsheet jockey. Question is, can the above type of data interaction scenarios be managed by someone who isn't a spreadsheet jockey or data scientist?
That's what I'm not sure about. It's becoming increasingly easy to "throw" a data analysis or visualization tool at a data set without really first understanding the data. No doubt that can be good as data analysis tools become simpler to use and more powerful. Exploratory analysis of any data set at minimum familiarizes the analyst with the data. It also helps refine and re-target the analysis process itself. That's also a good thing.
Still, will making data more amenable to analysis by a broader range of users, some of whom may require more analytical "hand-holding" than others, really help more people to ask and answer the questions important to them?
When I look at the success of Pokémon GO, I can't help but think that rising expectations about interactivity will raise receptivity -- and demand for -- more sophisticated and easy-to-use data analysis and visualization tools.
One possible application of the above thinking is the gamification of data analysis training, i.e., making it "fun" to learn how to extract meaning from an initially dirty and unorganized data set. Another possibility would be to make it easier to collaborate when analyzing data.
One idea that intrigues me is not the Tom Cruise/Minority Report "tactile" interface but rather the ability to walk around in and manipulate data directly. The model that comes to mind for this is how David from Prometheus walks inside the alien ship’s 3D navigational environment.
Still, taking responsibility for interacting directly with data is not an insignificant undertaking whether it's crunching numbers via a complex spreadsheet or using voice commands to instruct a system on analysis requirements. One needs a basic understanding of data and data analysis just to ask the right questions; can the system supply that understanding, or is it up to us?
Copyright (c) 2016 by Dennis McDonald
Published September 21, 2016, revised September 22.
The author would like to thank Jim Lola for providing thoughtful feedback on an earlier version of this paper.