Mining and modelling temporal dynamics of followers' engagement on online social networks

A relevant fraction of human interactions occurs on online social networks. In this context, the freshness of content plays an important role, with content popularity rapidly vanishing over time. We therefore investigate how influencers' generated content (i.e., posts) attracts interactions, measured by the number of likes or reactions. We analyse the activity of influencers and followers over more than 5 years, focusing on two popular social networks: Facebook and Instagram, including more than 13 billion interactions and about 4 million posts. We investigate the influencers' and followers' behaviour over time, characterising the arrival process of interactions during the lifetime of posts, which are typically short-lived. After finding the factors playing a crucial role in the post popularity dynamics, we propose an analytical model for the user interactions. We tune the parameters of the model based on the past behaviour observed for each given influencer, discovering that fitted parameters are pretty similar across different influencers and social networks. We validate our model using experimental data and effectively apply the model to perform early prediction of post popularity, showing considerable improvements over a simpler baseline.