12 years ago
Want to Improve Your Prototype-to-Production Analytics Process? Embrace Thinking Inside the Box

 Anyone in the business of analytics knows that the work is often highly iterative, exploratory and ill-defined. In my experience, as hard as it is to fit models that let us understand complex relationships and make predictions, it is often harder still to get such models into production environments. By “production environments” I mean settings where analytics are embedded in a business’s processes, as opposed to living only as pretty graphs in a PowerPoint. The business of analytics also means getting to play with new math and software tools when tackling new problems. I like to do pre-mortems