Deloitte published a fascinating annual trends report named Tech Trends 2013, which includes several chapters of insightful and forward-thinking discussions on where technology is headed. What caught my eye are the chapters on the role of CIOs and of finding ways to understand Big Data. I’ll reserve the CIO topic for later. Instead, let’s discuss Big Data. Or, actually, let’ use Big Data as a segue into how tough problems like analyzing Big Data can help even “small data” companies like mine, and yours, better understand the information in front of them.
(Oh, and yes, “Big Data” is usually capitalized. It’s industry jargon.)
Big Data is, well, lots of data. Terabytes of it. Companies like Facebook, Craigslist, and Twitter deal with Big Data every day. The government has to process reams of it. IT departments for organizations like Boeing store it in data warehouses so they can view, pivot, slice, dice, and chop it up into understandable patterns.
And they all have the same problems: How to understand Big Data. And it’s a Big Problem because THERE IS SO MUCH OF IT, and so it’s not only tough to manage the sheer weight of data, but also how to interpret it.
Typical approaches to interpreting the data include bar graphs and pie charts, KPIs and reports, and an assortment of other techniques. What I see happening in the near-future is the expanded use of Chernoff faces and related techniques.
I first heard about Chernoff faces while listening to, if memory serves, an episode of Science Friday on NPR. (At least I’m pretty sure it was Science Friday—my reminder in my ‘To Blog About’ list says “Chernoff faces… NPR”.) Chernoff faces work like this: You take something the human brain is very good at—understanding and remembering faces—and use that innate mental strength to bridge the gap between our limited ability to see patterns in raw numbers and data and our ability to easily find patterns in visual representations. In this case, human faces.
Take a quick look at a representation, via human faces, of how certain judges are known to rule. Now imagine trying to figure that out by reading though case studies or even numeric data. It’s a lot easier to see patterns via faces, right?
The cool thing about Chernoff faces is that this isn’t specific to Big Data. It can just as easily be used to represent Small Data. For example, if you are in Marketing and want to get a good understanding of your customer base and their experience with you, you could perform a survey with a set number of questions and available ratings, define each question as a part of a face (e.g., a customer’s willingness to refer you to another person is assigned the eyebrows), and then define the rating as how that part of the face is drawn (e.g., happy, angry, sad each indicate a rating of 3, 2, and 1).
You’ll get an idea pretty quickly of where your customer base stands. Now imagine having a dashboard showing your top 10 customers as Chernoff faces and being able to see the faces change over a period of, say, a year’s worth of surveys. Your mind will naturally determine your customer satisfaction trending.
Mix a little psychology, natural human strengths, and technology, and you can develop a system to understand important data very quickly. Try the technique out and let me know how it goes.