Efficiency of human activity on information spreading on Twitter

• Published in: Social networks Vol. 39; pp. 1 - 11
• Main Authors: Morales, A.J., Borondo, J., Losada, J.C., Benito, R.M.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.10.2014
• Subjects: Collective action; Complex networks; Individual Collective Relationship; Influence; Information spreading; Mass Media Effects; Messages; Networks; Quantitative analysis; Quantitative Methods; Social media; Social networks analysis; Topology; Twitter; User behavior; Twitter; Influence; Information spreading; User behavior; Complex networks; Social networks analysis
• ISSN: 0378-8733; 1879-2111
• DOI: 10.1016/j.socnet.2014.03.007

•Define user efficiency as the rate between influence gained and efforts spent.•Diverse Twitter datasets show universal behavior in the user efficiency patterns.•Propose a spreading model to explore the effects of the topology and user behavior.•Underlying networks heterogeneity determine the emergence of highly efficient users.•Activity itself cannot enhance nor compensate the individual topological deficits. Understanding the collective reaction to individual actions is key to effectively spread information in social media. In this work we define efficiency on Twitter, as the ratio between the emergent spreading process and the activity employed by the user. We characterize this property by means of a quantitative analysis of the structural and dynamical patterns emergent from human interactions, and show it to be universal across several Twitter conversations. We found that some influential users efficiently cause remarkable collective reactions by each message sent, while the majority of users must employ extremely larger efforts to reach similar effects. Next we propose a model that reproduces the retweet cascades occurring on Twitter to explain the emergent distribution of the user efficiency. The model shows that the dynamical patterns of the conversations are strongly conditioned by the topology of the underlying network. We conclude that the appearance of a small fraction of extremely efficient users results from the heterogeneity of the followers network and independently of the individual user behavior.