Characterizing user behavior in online social networks

By Fabrício Benevenuto, Tiago Rodrigues, Meeyoung Cha, Virgílio A. F. Almeida. ACM/SIGCOMM Internet Measurement Conference, 2009

904 citations recorded by the source corpus.

Understanding how users behave when they connect to social networking sites creates opportunities for better interface design, richer studies of social interactions, and improved design of content distribution systems. In this paper, we present a first of a kind analysis of user workloads in online social networks. Our study is based on detailed clickstream data, collected over a 12-day period, summarizing HTTP sessions of 37,024 users who accessed four popular social networks: Orkut, MySpace, Hi5, and LinkedIn. The data were collected from a social network aggregator website in Brazil, which enables users to connect to multiple social networks with a single authentication. Our analysis of the clickstream data reveals key features of the social network workloads, such as how frequently people connect to social networks and for how long, as well as the types and sequences of activities that users conduct on these sites. Additionally, we crawled the social network topology of Orkut, so that we could analyze user interaction data in light of the social graph. Our data analysis suggests insights into how users interact with friends in Orkut, such as how frequently users visit their friends' or non-immediate friends' pages. In summary, our analysis demonstrates the power of using clickstream data in identifying patterns in social network workloads and social interactions. Our analysis shows that browsing, which cannot be inferred from crawling publicly available data, accounts for 92% of all user activities. Consequently, compared to using only crawled data, considering silent interactions like browsing friends' pages increases the measured level of interaction among users.

Subfields: Computer Networks and Communications, Statistical and Nonlinear Physics

View Paper: https://www.semanticscholar.org/paper/9bd3c0cbdd7c73e67f39abb9dd5fc3714b0965a5

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Statistical and Nonlinear Physics

Citations in the corpus

What is Twitter, a social network or a news media? (2010, background): “Online social networks and social media The rising popularity of online social networking services has spurred research into their characteristics and recent work has forayed into characteristics beyond crawled data [ 3 , 39].”
The little engine(s) that could: scaling online social networks (2010, background): “…Pablo Rodriguez Telefonica Research { jmps, vijay, georgos, yxiao, nikos, pchhabra, pablorr }@tid.es ABSTRACT The di.culty of scaling Online Social Networks (OSNs) has introduced new system design challenges that has often caused costly re-architecting for services like Twitter and Facebook.”
Combating Attacks and Abuse in Large Online Communities (2016, methodology): “As in pr ior work, we assume that a user’s session is over if they do not make any requests for 20 minutes [47].”
User Migration in Online Social Networks: A Case Study on Reddit During a Period of Community Unrest (2016, background): “Many users do not comment and fewer still create posts; these users represent a large silent majority, a phenomenon which is observed on many other types of social media (Benevenuto et al. 2009; Bernstein et al. 2013; Gong, Lim, and Zhu 2015).”
#mytweet via Instagram: Exploring user behaviour across multiple social networks (2015, background): “To understand how user interact with multiple OSNs, early studies exploited user clickstream data from passively monitored network traffic [8] or a social network aggregator [9].”
Understanding latent interactions in online social networks (2010, background): “Recently, two signiﬁcant studies [3, 25] used clickstream data at the network level to capture the behavior of OSN users, and revealed that passive or latent interactions such as proﬁle browsing often dominate user events in a social network [3].”
Making Geo-Replicated Systems Fast as Possible, Consistent when Necessary (2012, neutral): “[4].”
Understanding online social network usage from a network perspective (2009, background): “To analyze such properties one has to capture the interactions of the user with the OSN over time which is impossible via crawling.”
Uncovering social network sybils in the wild (2011, methodology): “As in prior work, we define a session as over when a user does not click any links for 20 minutes [Benevenuto et al. 2009].”
Understanding Short-term Changes in Online Activity Sessions (2017, background): “Benevenuto et al. have aggregated data from multiple social networks and created sessions of activity to understand high-level behavior of users in usage of online social networks, e.g., how frequently people login to these sites and how much time they spend browsing them [1, 2].”
Understanding user behavior in online social networks: a survey (2013, methodology): “Benevenuto et al. [6] analyze the user behavior of OSNs based on detailed clickstream data obtained from a social network aggregator, as illustrated in Fig.”
Characterizing Silent Users in Social Media Communities (2015, background): “In practice, however, lurking is a very common behavior found in many content providing sites (Nonnecke and Preece 2000; Muller et al. 2010; Antin and Cheshire 2010; Benevenuto et al. 2009; Bernstein et al. 2013).”
Detecting and characterizing social spam campaigns (2010, background): “studied how users interact with friends, like how frequently users visit their friends’ pages [18].”
Minimax Rates of Entropy Estimation on Large Alphabets via Best Polynomial Approximation (2014, background): “…(3) and (5), we reach the following verdict on the plug-in estimator: In the regime where the sample size n grows at most polynomially in k , empirical entropy is rate-optimal, i.e., within a constant factor of the minimax risk, if and only if we are in the “data-rich” regime n = Θ( k 2 log 2 k ).”