Crawling facebook: a social network analysis

Abstract

Online Social Networks (OSN) play an integral role in our everyday life, affecting the social life and activity of people in various ways. Social Networking sites have hundreds of millions of registered users who use these sites to share thoughts, experiences, photographs, meet new people, contact long-lost friends and family members, find jobs, spread information, and more The idea of social networks, and that social phenomena can be explained when we surpass the properties of individuals and examine their personal and social ties, has been around for over a century. Social Networks play a critical role in the social, economic, health, educational aspects of our life and behavior in general. Their structure affects the way information flows amongst people, the way diseases spread, our purchase choices, the decisions we make and the way our society evolves. In this Thesis we perform a study that includes crawling the most popular online social network site "Facebook" and performing a proof-of-concept Social Network Analysis. We describe the collection process of the crawlers implemented in python. Moreover we provide graph visualization and study several graph metrics with the help of Gephi, an open source program for visualizing and analyzing large graphs. We provide metrics and analyze network graph properties such as degree distribution, centrality measures, and community detection, among others. From our extracted anonymized data we choose to further analyze users’ likes in conjunction with their relationships and provide basic statistics and analysis. We analyze the community detection mechanism and raise the question if community unfolding results can be reproduced and/ or improved or if we take into consideration the users common preferences (likes).