Analysis of Topological Characteristics of Huge Online Social Networking Services

Analysis of Topological Characteristics of Huge Online Social Networking Services

WWW 2007 / Track: Semantic Web Session: Semantic Web and Web 2.0 Analysis of Topological Characteristics of Huge Online Social Networking Services Yong­Yeol Ahn Seungyeop Han Haewoon Kwak Department of Physics Division of Computer Science Division of Computer Science KAIST, Deajeon, Korea KAIST, Daejeon, Korea KAIST, Daejeon, Korea [email protected] [email protected] [email protected] Sue Moon Hawoong Jeong Division of Computer Science Department of Physics KAIST, Daejeon, Korea KAIST, Deajeon, Korea [email protected] [email protected] ABSTRACT million users 2 years ago, one fourth of the entire popula- Social networking services are a fast-growing business in the tion of South Korea. MySpace and orkut, similar social net- Internet. However, it is unknown if online relationships and working services, have also more than 10 million users each. their growth patterns are the same as in real-life social net- Recently, the number of MySpace users exceeded 80 million works. In this paper, we compare the structures of three with a growing rate of over a hundred thousand people per online social networking services: Cyworld, MySpace, and day. It is reported that these SNSs “attract nearly half of all orkut, each with more than 10 million users, respectively. web users” [1]. The goal of these services is to help people We have access to complete data of Cyworld’s ilchon (friend) establish an online presence and build social networks; and relationships and analyze its degree distribution, clustering to eventually exploit the user base for commercial purposes. property, degree correlation, and evolution over time. We Thus the statistics and dynamics of these online social net- also use Cyworld data to evaluate the validity of snowball works are of tremendous importance to social networking sampling method, which we use to crawl and obtain par- service providers and those interested in online commerce. tial network topologies of MySpace and orkut. Cyworld, The notion of a network structure in social relations dates the oldest of the three, demonstrates a changing scaling be- back about half a century. Yet, the focus of most sociological havior over time in degree distribution. The latest Cyworld studies has been interactions in small groups, not structures data’s degree distribution exhibits a multi-scaling behavior, of large and extensive networks. Difficulty in obtaining large while those of MySpace and orkut have simple scaling be- data sets was one reason behind the lack of structural study. haviors with different exponents. Very interestingly, each However, as reported in [2] recently, missing data may dis- of the two exponents corresponds to the different segments tort the statistics severely and it is imperative to use large in Cyworld’s degree distribution. Certain online social net- data sets in network structure analysis. working services encourage online activities that cannot be It is only very recently that we have seen research re- easily copied in real life; we show that they deviate from sults from large networks. Novel network structures from close-knit online social networks which show a similar de- human societies and communication systems have been un- gree correlation pattern to real-life social networks. veiled; just to name a few are the Internet and WWW [3] and the patents, Autonomous Systems (AS), and affiliation net- Categories and Subject Descriptors: J.4 [Computer works [4]. Even in the short history of the Internet, SNSs are Applications]: Social and behavioral sciences a fairly new phenomenon and their network structures are General Terms: Human factors, Measurement not yet studied carefully. The social networks of SNSs are Keywords: Sampling, Social network believed to reflect the real-life social relationships of people more accurately than any other online networks. Moreover, because of their size, they offer an unprecedented opportu- 1. INTRODUCTION nity to study human social networks. The Internet has been a vessel to expand our social net- In this paper, we pose and answer the following questions: works in many ways. Social networking services (SNSs) are What are the main characteristics of online social net- one successful example of such a role. SNSs provide an on- works? Ever since the scale-free nature of the World-Wide line private space for individuals and tools for interacting Web network has been revealed, a large number of networks with other people in the Internet. SNSs help people find have been analyzed and found to have power-law scaling in others of a common interest, establish a forum for discus- degree distribution, large clustering coefficients, and a small sion, exchange photos and personal news, and many more. mean degrees of separation (so called the small-world phe- Cyworld, the largest SNS in South Korea, had already 10 nomenon). The networks we are interested in this work are huge and those of this magnitude have not yet been ana- Copyright is held by the International World Wide Web Conference Com­ lyzed. mittee (IW3C2). Distribution of these papers is limited to classroom use, How representive is a sample network? and personal use by others. In most networks, WWW 2007, May 8–12, 2007, Banff, Alberta, Canada. the analysis of the entire network is infeasible and sampling ACM 978­1­59593­654­7/07/0005. 835 WWW 2007 / Track: Semantic Web Session: Semantic Web and Web 2.0 is unavoidable. We evaluate the popular snowball sampling 3. ONLINE SNSS method using the complete Cyworld network topology. Social networking services (SNSs) provide users with an How does a social network evolve? From the historical online presence that contains shareable personal informa- data of the Cyworld network, we reconstruct the past snap- tion, such as a birthday, hobbies, preferences, photographs, shots of Cyworld network. The evolution path of a social writings, etc. Most SNSs offer features of convenience that network may exhibit patterns of major change. It would help users form and maintain an online network with other be of tremendous interest to latecomers, for it is literally a users. One such feature is a “friend.” A user invites an- prediction of what they might turn into. other user to be one’s friend. If the invited user accepts This paper is organized as follows. In Section 2, we re- the invitation, a friend relationship is established between view related works. In Section 3, we describe social network them. This friend feature is often used as a shortcut to topologies we have gained access to and, in Section 4, the others’ front pages and, in some SNSs, a convenient tool to snowball sampling method and metrics of interest in net- exchange messages and stay in touch. work analysis. Then we begin our topology analysis with As SNSs have become very popular, leading sites boast of that of Cyworld. In Section 5, we analyze the friend rela- an extensive user base of tens of millions of subscribers. For tionship network of Cyworld, perform a historical analysis this work, we have collected four sets of online social network on the network’s evolution, evaluate the snowball sampling topology data from three popular SNSs. All of them have method, and compare a special subset of Cyworld network more than 10 million users. We have obtained the entire (the network of testimonials) with the complete network, network topology of Cyworld directly from its provider, and and . In Section 6, we analyze MySpace and orkut net- sampled others through web crawling. We believe our work works. We compare all three networks in Section 7, along is the first to analyze social networks of such magnitude. In with a discussion on the origin of power-law in online social Table 1, we summarize the four data sets described above. networks and the resemblance to real social networks. We The metrics in the table are explained later in Section 4. conclude with a discussion on future work in Section 8. Below we describe each network in detail. 2. RELATED WORKS 3.1 Cyworld The structural propertes, such as degree distribution, of Cyworld is the largest and oldest online social networking large-scale social networks have received much attention since service in South Korea. It began operation in September the uncovering of the network of movie actors [3]. It is fol- 2001, and its growth has been explosive ever since. Cy- lowed by the analysis on the network of scientific collabora- world’s 12 million registered users, as of November 2005, tion network [5] and the web of human sexual contacts [6]. are an impressive number, considering the total population However, a link in these networks is different from a normal of 48 million in South Korea. As any SNS, Cyworld offers friend relationship and the large-scale analysis on the such users to establish, maintain and dissolve a friend (called il- networks has remained uncharted. chon) relationship online. Recently, the rapid growth of online social networking ser- From SK Communications, Inc. the provider of the Cy- vices made it possible to investigate the huge online social world service, we received an anonymized snapshot of the network directly. Since the rise of Cyworld, many SNSs in- Cyworld user network taken in November 2005. The snap- cluding MySpace and orkut have grown. However, the anal- shot contains 191 million friend relationships among 12 mil- yses on these huge networks have been limited to cultural lion users. We have access to additional data to study the and business viewpoint [7]. network evolution and describe it in Section 5.2. Cyworld offers a mechanism called ilchon pyung for friends Here, we introduce two relevant works on online social 1 networks. First work is on an Internet dating community, to leave a testimonial on a user’s front page Friends can called pussokram.com [8].

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    10 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us