Analysis of Proximity Informed User Behavior in a Global Online Social Network

TL;DR

This paper investigates how physical distance influences online social connections globally, revealing country-specific patterns and stronger effects when in-person meetings are possible, with differences between weak and strong ties.

Contribution

It introduces a dyadic Logit model to analyze spatial dependence in online social networks, highlighting country-specific and tie-strength variations in proximity effects.

Findings

Distance effects vary by country

Proximity influences are stronger with potential in-person meetings

Weak ties are more dependent on proximity than strong ties

Abstract

Despite the earlier claim of "Death of Distance", recent studies revealed that geographical proximity still greatly influences link formation in online social networks. However, it is unclear how physical distances are intertwined with users' online behaviors in a virtual world. We study the role of spatial dependence on a global online social network with a dyadic Logit model. Results show country-specific patterns for distance effect on probabilities to build connections. Effects are stronger when the possibility for two people to meet in person exists. Relative to weak ties, dependence on proximity is looser for strong social ties.