While data mining can unearth a wealth of information, it takes discriminating analysis to make sure we are not just making connections, but the right connections.

I had a discussion recently with someone who, for three decades, had been performing a statistical function at a large manufacturing company. He couldn’t understand why, in spite of excellent job performance reviews, his company had furloughed him indefinitely.

As the story goes, in an effort to improve sales and efficiency, a grocery store conducted a study to see which items were most commonly purchased together by its patrons.