AAPOR gets it wrong
Aligning practice and principle in the NewMR

"Just trust us."

While skimming an article in the current issue of Businessweek on the Weather Channel’s emerging use of big data for things other than predicting the weather, I saw the graphic at the right and immediately thought of the causality versus correlation debate.  Feat_weatherchart42b_630They have gotten hold of tons of data from Walmart and P&G, which they are putting together with local weather data. Then they make predictions like these.

I was intrigued enough to look at the text more closely and found this from an interview with Vikram Somaya, who seems to sit on top of their big data operations:

“We can tell you that on a January morning in Miami, if a set of weather conditions occurs, people will buy a certain brand of raspberry,” he says. Not just any fruit. Raspberries. When advertisers ask for an explanation—why raspberries?—Somaya can’t always provide a clear answer. “A lot of times we have to tell them to just trust us.

This from a company whose core business is predicting the weather. The world is changing faster than I have imagined.