A few years ago our new media team made this quick video to give you sense of just how big a billion is:
But now that you can visualize a billion dollars, can you imagine what a billion rows of data looks like? Or find value in those billions of variables? Not really. And that’s just one reason why data visualization is important for “big data.” Most of us can’t understand what we can’t see.
On the other hand, if you can visualize billions and billions of values by quickly distilling them into subsets, graphs and box plots, you can start to understand the data better. Essentially, visualization gives analysts a jump start on the modeling processes by making it easy to explore the relationships between variables and attributes. When you can actually see and query that much data, you can start to determine which fields mean something and which you should maybe discard.
Before high-performance analytics, you could only do these types of explorations on subsets of large data sets. Now, high-performance analytics make it possible to visualize all 5 billion (plus) rows of data.
What’s the significance? Whether it’s 1 billion or 5 billion rows of data, in-memory analytics is changing the way organizations look at and model data. This becomes even more important when viewed as part of a larger process. Discovery from the visualizations can be fed into more in-memory processes where advanced analytics, like marketing optimization, are applied to improve the business.
With all technology applications, the real benefit will come when high-performance processes are put in to context for business solutions, such as merchandise planning, assortment planning, fraud detection, and value-at-risk calculations. The visualizations are exciting and can provide the early insights, but don't forget: improving business processes is the essential next step.