Abstract:
Public profiles of social network users can be a source of demographic and sociological
information, in particular providing data on the intensity of interregional interaction («region friendship»). Previously user profiles from the VK social network were indexed and
processed to form a geotagged database, published within the webcensus.ru project. The
present article uses this data to discover relationships between intensities of interregional
migration and friendship for municipal districts and regions of Russia. We investigate
four models for estimating the amount of friend links between regions using data on the
number of migrants between those regions. We show that the model that takes into account the entire system of interregional migration flows is better at predicting the number
of friendship links than models which only consider every pair of interacting regions
separately from the others.
Keywords:social network, region interaction, migration, graph model.