Speaking at the Hadoop Summit on Wednesday, Netflix Senior Data Scientist Mohammad Sabah revealed that the streaming media giant is gathering and analyzing an incredible amount of data to power its recommendation system. Sabah says that 75 percent of users select movies based on the company’s recommendations.
So what is Netflix collecting? GigaOm provides a nice list. Some highlights:
-25 million+ users
-30 million plays per day
-2 billion+ hours of streaming video watched during Q4 2011
-4 million ratings per day
-3 million searches per day
Netflix also collects geolocation, device, and time-of-day data, in addition to metadata from third parties like Nielsen, Facebook and Twitter.