Sociolinguistic Journal of Korea, The purpose of this paper is to explore the influential effect of social media interaction on English learning in terms of socio-cultural perspectives. The data were collected from two groups of subjects (A:38 B:40) taught by native English teachers and were analyzed by Nvivo 10 and NodeXL for the detailed discourse research between interlocutors. Halliday(1975)`s Model of Language Function was adopted for the initial analysis, and afterwards worked well with the influential facdtors of social media as described in Aaker & Smith(2010)`s dragonfly effect. The result shows that social media features made students` RF (responsive feedback) and teachers` CF (corrective feedback) rapid, which in turn lead to very active communication. Furthermore, the communication was of various types, such as explicit correction, recasting, clarification repuesting, meta linguistic response, repetition, elicitation, circumlocution, and asking for assistance. In the data analysis done using NodeXL, the analytic induction suggests that the reaction in social media had a high betweeness centrality score, and the centrally hitched cloud-type interaction constitutes 92% of participants` interaction, which means most students were involved in the interaction. The discourse analysis also illustrates the influential impact on socio-cultural interaction of social media. Overall, the potential possibility of using social media for English education is shown to have strong and positive implications.