Adult content consumption in online social networks

• Published in: Social network analysis and mining 2017-12, Vol.7 (1), p.28, Article 28
• Main Authors: Coletto, Mauro, Aiello, Luca Maria, Lucchese, Claudio, Silvestri, Fabrizio
• Format: Article
• Language: English
• Subjects: Age; Applications of Graph Theory and Complex Networks; Behavior; Community; Computer Science; Consumption; Data Mining and Knowledge Discovery; Datasets; Demography; Digital media; Drug abuse; Eating disorders; Economics; Females; Game Theory; Gender; Humanities; Inappropriateness; Law; Methodology of the Social Sciences; Network analysis; Original Article; Pornography; Pornography & obscenity; Social and Behav. Sciences; Social media; Social networks; Social sciences; Statistics for Social Sciences; Subgroups; Virtual communities
• ISSN: 1869-5450; 1869-5469
• DOI: 10.1007/s13278-017-0449-y

Users in online social networks naturally organize themselves into overlapping and interlinked communities that are formed around common identity or shared topical interests. Some communities gather people around specific deviant behaviors , conducts that are commonly considered inappropriate with respect to the society’s norms or moral standards such as drug use, eating disorders, and pornographic content consumption. From a network analysis perspective, the set of interactions between members of these communities form deviant networks that map how the deviant content is shared and consumed. It is commonly believed that deviant networks are small and isolated from the mainstream social media life; accordingly, most research studies have considered them in isolation. We focus on adult content consumption networks, which is one deviant network with a significant presence in online social media and in the Web in general. We investigate two large online social networks and discuss the following insights. Deviant networks are limited in size, tightly connected and structured in subgroups. Nevertheless, content originated in deviant networks spreads widely across the whole social graph possibly touching a large number of unintentionally exposed users, such that the average local perception is that neighboring users share more deviant content. Finally, we investigate how content production and consumption vary with age and show that the consumption rate is very similar between male and female users up to the age of 25. We conclude that deviant communities are deeply rooted into the relational fabric of a social network, and that a deeper understanding of how their activity impacts on every other user is required.