Wednesday, February 24, 2016

"Analytics Reality: Do You Excel?"

Lifted in toto from Beyond Search:
I read “What Is the Most used Feature in Any business Intelligence Solution? It’s the Export to Excel Button.” The write up asserts:
I was recently forwarded an article on the continued popularity of Excel in the BI community consisting of quotes from 27 experts saying how great and how relevant Excel remains. We do categorize BI as static and historical as opposed to forward looking predictive analytics but I bet it’s still true that Excel is a very widely used tool even by folks that categorize themselves as data scientists.
Let’s assume this is accurate. What does this suggest for complex analytics like my old pals SAS or IBM SPSS? What about high flying outfits like Palantir, and Centrifuge Systems?

I have some answers, but I think the questions are suggestive of a hurdle which high horsepower analytic systems must power around. There is a reason so few folks are adept at statistics whether the industrial strength variety or the weird approach taken in social science and economics classes.
Excel seems to be tough to master but compared to more supercharged methods, Excel sure looks like a push peddle tricycle. You can’t go too far or too fast. If you crash into something, there is F1 and semi automated procedures to kiss the boo boo and make it better.
Stephen E Arnold, February 24, 2016