Role discovery in node-attributed public transportation networks: the study of Saint Petersburg city open data

The work presents results of modeling Public Transportation Networks (PTNs) of Saint Petersburg (Russia) and highlights the roles of stations (stops) in this network. PTNs are modeled using a new approach, previously proposed by the authors, based on weighted networks with node attributes. The nodes correspond to stations (stops) of public transport, grouped according to their geospatial location, while the node attributes contain information about social infrastructure around the stations. Weighted links integrate information about the distance and number of transfers in the routes between the stations. The role discovery is carried out by clustering the stations according to their topological and semantic attributes. The paper proposes a software framework for solving the problem of discovering roles in a PTNs. The results of its application are demonstrated on a new set of data about the PTNs of Saint Petersburg (Russia). The significant roles of the nodes of the specified PTNs were discovered in terms of both topological and infrastructural features. The overall effectiveness of the PTNs was assessed. The revealed transportation and infrastructural shortcomings of the PTNs of Saint Petersburg can be considered by the city administration to improve the functioning of these networks in the future.