Tracking the dynamic variations in a social network formed through shared interests

Abstract

We tracked the dynamics of a social network formed by a shared interest in movies. Users-, movie ratings-, and rental date-data from the Netflix Prize dataset were used to construct a series of date-filtered social networks, wherein viewers were linked when they rented the same movie and gave the same rating. We obtained a nearly constant high clustering coefficient (0.60 – 0.85), and a low average path length (1.4 – 2.3) indicating a static 'small-world' network despite the dynamic behavior of the borrowers.

Keywords

• networks
• time series analysis
• complex systems
• small-world network