New Model Captures How Multiple Social Traits Shape Network Connections

Homophily, the tendency of people to connect with others who resemble them, is one of the most powerful forces organizing social networks. A new study in EPJ Data Science argues that this force has too often been modeled in a simplified way, treating social similarity as if it depended on a single attribute rather than the many overlapping traits that define real social identities.

Researchers Rubén Rodríguez-Casañ, Javier Borge-Holthoefer and Alessio Cardillo introduce a multidimensional extension of the Social Distance Attachment model. In their framework, individuals are represented as vectors in a metric space, where each dimension corresponds to a social feature. Crucially, the model distinguishes between exclusive features, such as gender, and non-exclusive features, such as language or thematic specialization, allowing both types to coexist within a single distance-based description of social proximity.

The key innovation is that different dimensions can contribute unequally to network formation. Rather than assuming every attribute matters equally, the model lets researchers assign different weights to each dimension, reflecting the possibility that the relative importance of social traits varies across contexts.

To test the framework, the team applied it to two empirical networks of cultural production. The second is a network of contemporary global music collaborations drawn from Spotify data, connecting artists through their joint work.

The results indicate that the interaction among social dimensions plays a central role in shaping how these networks are organized.

The model also produced a more nuanced picture of how heterophily, the tendency to connect with dissimilar others, affects minority groups. The researchers found that the effects of heterophily on minority centrality depend critically on correlations between attributes and on the weighting assigned to each dimension. In other words, whether belonging to a minority helps or hinders a person’s position in a network is not a fixed property but emerges from the configuration of the whole multidimensional space.

These findings underscore why multidimensional representations of social profiles matter for understanding homophily-driven network formation. A model that collapses identity into one dimension risks misreading both the structure of communities and the position of individuals within them, particularly when attributes are correlated with one another.

Beyond explaining existing networks, the authors position the framework as a flexible tool for exploring inclusion dynamics in complex social systems. By adjusting which dimensions are emphasized and how they are weighted, researchers can probe how different social configurations might shape access to central positions in a network.

The study is published as open access research in EPJ Data Science, with the article released under a citable version carrying a permanent DOI.

Subject of Research: Multidimensional modeling of homophily in social networks

Article Title: Modeling social networks with homophily via multi-dimensional social distance attachment

Article References: Rodríguez-Casañ, R., Borge-Holthoefer, J., & Cardillo, A. (2026). Modeling social networks with homophily via multi-dimensional social distance attachment. EPJ Data Science. https://doi.org/10.1140/epjds/s13688-026-00713-7

Image Credits: AI Generated

DOI: 10.1140/epjds/s13688-026-00713-7

Keywords: homophily, social networks, network science, Social Distance Attachment, multidimensional modeling, minority centrality, community structure, heterophily, music collaborations, intellectual cooperation, Spotify, EPJ Data Science