Here’s to more questions – and answers – in 2013

I’m sensing a theme as I look over my posts on this blog from the last year. The top five posts give you a good taste of what we were thinking about as an industry this year:

  1. Is big data over hyped?
  2. What kind of big data problem do you have?
  3. What is a data scientist? And do you really need one?
  4. SAS = analytics. Analytics = hot.
  5. Why would you want to visualize 5 billion rows of data?

To summarize: big data, big data, data science, big analytics, and big data visualization.

My goal in writing about these topics was to pull away from the industry hype and offer some real definitions and explanations for what these trends mean for your business.

If you’ve been reading along, did I succeed? Are you bought into the idea of finding real, forward-looking answers in your new big data sources?

Have we talked enough, or are we ready to get going in 2013? Once we make it through the holidays and survive the doomsday predictions, are you ready to visualize your big data?

This quote from a new visual analytics user at SAS really sums up the benefits: “You can’t look through five-and-a-half million lines of data and make sense of anything. But you can with Visual Analytics.”

Or, how about this one from XL Group VP Kimberly Holmes: "Visual analytics will inspire more questions than we ever would have asked before."

Join me here next year, and we’ll talk more about the types of questions you should be asking, and the types of answers you’re getting – from your big data.

tags: big analytics, big data, data visualization, high-performance analytics

One Comment

  1. Ken Comer
    Posted February 1, 2013 at 10:58 am | Permalink

    Well, I'm sad that I'm the first to add a comment, and it's a month after your article! This is really something we need to be discussing. I sense, from your posts in 2012, Jim, that you are interested in an objective discussion of how big data and big analytics (and their companions) should be viewed by non-quantitative business leaders.

    As I noted in my reaction to your February 2012 article, the acceptance of the 'big data paradigm' has been nearly universal. Yours is the only mildly 'critical' thought series that shows up in Google.

    It seems as if you're asking for a discussion as I re-read your article, so I decided it would be useful to hit the tennis ball back.

    Let me offer my reactions as a former SAS customer and current operations research professional.

    There are underlying assumptions here that should be aired and examined by somebody considering an expanded effort in the big data arena. Just because we can execute predictive algorithms against these oversized data sets does not mean that the predictions will be accurate, insightful, or drive us to the right decisions.

    In the risk arena, I would ask some probing questions. What predictions have been made successfully with this tool ('detections'); what important risk events have not been made with this tool ('misses'); and what predictions have been made that have not come true ('false alarms'). Most importantly, where does a SAS analytic process operate in the trade space between misses and false alarms. Everybody must make that tradeoff, what has been the SAS history?

    In the fraud detection arena, while I recognize SAS's impressive ability to detect some fraud, I would wonder if there are any extrinsic indications of the scale of the fraud that is not being detected. Consider the Lance Armstrong doping problem: He won numerous Tours under intense scrutiny and massive testing. Apparently, he and most of his team's doping went undetected by the testing process for years. We sure didn't get our money's worth from that testing program! Yet, at the time the doping tests could show 'success' in detecting several individuals. It appeared, for a time, as a successful 'fraud detection' system. This is the problem when you use 'detections' (alone) as your MOE for a surveillance program.

    I make these comments in the hope that the SAS sales team adjust its focus. I think a much more convincing argument for a big data investment can be made by describing the changes in outcomes versus better 'insights'. I doubt that business leaders want to be inspired just to ask more questions. They want to have some confidence they are driving toward the best decisions using rational processes that have stood the test of time.

Post a Comment

Your email is never published nor shared. Required fields are marked *

*
*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <p> <pre lang="" line="" escaped=""> <q cite=""> <strike> <strong>