Big Data is a big problem. How do you distill useful meaning from it?
“Amazon Mines Its Data Trove To Bet on TV’s Next Hit,” Wall Street Journal, November 2, 2013 A1 http://on.wsj.com/1hCMGdB
Amazon picks which TV shows it will put on streaming video by putting 14 “pilots” on the web, letting a million people view them, and harvesting how much each pilot was viewed, how many viewers gave a video 5 stars, and how many viewers shared it with friends.
How do you pick the “metrics” to measure? There is apparently no objective definition of “good” or “great” when it comes to TV shows. Letting the crowd decide, as opposed to some “tastemaker” in Hollywood. Hollywood tests its proposed shows, but with as few as 50 people. Apparently sample size matters.
And this raises the question of filters. Don’t the people who select what shows make it to market act as filters? Is that a good thing or a bad thing? They keep us from having to sort through thousands of options to find what we really want. So what if they exclude some things that we would have liked?