Clustering Foursquare Mobility Networks to Explore Urban Spaces

Our study aimed to explore Foursquare mobility networks and investigate phenomena of clustering venues across the cities. We performed graph-based clustering to detect venues that highly interact among each other in terms of aggregated users mobility flows. Available Foursquare data included check-in information for ten large worldwide cities, observed in the period of two years, each having large number of geo-tagged venues coupled with semantic information in form of venue category. Such data allowed us to study cities as complex systems and explore their dynamic nature. We obtain global overview on the semantics content of clusters derived from venues categories, quantified changes in the clusters on a monthly bases and compared results between cities.