In the history of recommendation algorithms, few datasets are as revered or as studied as the Netflix Prize dataset, which reached a critical fever pitch in 2007. For data scientists, the "index" of this dataset wasn't just a list of movies; it was a map of human behavior.