Beyond Loyalty: Mapping and Predicting Poly-Community User Archetypes Across Platforms

While Hamilton et al. (2017) focused on loyalty—users devoted predominantly to one community—there’s an entire spectrum of users who fluidly engage in several communities (“poly-community users”). With datasets like MADOC (Mitrovic et al., 2025), we can, for the first time at scale, cluster users by their engagement styles across platforms (e.g., “serial joiners,” “bridge-builders,” “opportunists,” “identity shifters”). This research would use machine learning and network analysis to define these archetypes, track their migration patterns, and predict their future engagement. The novelty lies in breaking away from single-community loyalty paradigms, leveraging cross-platform data to understand the dynamics and influence of users who weave the social fabric between communities. Such insights could transform moderation, recommendation, and community design strategies.

References:

1. Multi-Platform Aggregated Dataset of Online Communities (MADOC). Marija Mitrovic, Aleksandar Tomašević, S. Maletić, Miroslav Andjelković, Anja Vranic, Darja Cvetković, Boris Stupovski, Dusan Vudragovic, Sara Major, Aleksandar Bogojevi'c (2025). International Conference on Web and Social Media.
2. Loyalty in Online Communities. William L. Hamilton, Justine Zhang, Cristian Danescu-Niculescu-Mizil, Dan Jurafsky, J. Leskovec (2017). International Conference on Web and Social Media.
3. Multi-Platform Aggregated Dataset of Online Communities (MADOC). Marija Mitrovic, Aleksandar Tomašević, S. Maletić, Miroslav Andjelković, Anja Vranic, Darja Cvetković, Boris Stupovski, Dusan Vudragovic, Sara Major, Aleksandar Bogojevi'c (2025). International Conference on Web and Social Media.