The variability in population density in large cities during the day does not always lend itself to straightforward monitoring and control. This is particularly true for public spaces attracting tourists (embankments, parks, pedestrian zones etc.) and those not linked to professional or consumer activity of the city residents.
We can receive data on the attractiveness of various objects and territories based on the corresponding activity of users in social networks. To do this, we collect and analyze geographically tagged data (photos in the Instagram, Twitter posts), which generally allows you to build a picture of public space attendance. Further, with the use of the general model of mobility of the population (accounting for the sex and age structure as well as social stratification), the total number of people present in a given area is estimated given the related number of identified users of social networks. In general, this approach allows us to compensate the population data missed out by other sources of urban statistics and analyze the social network sites’ users perceptions of various urban facilities based on sentiment analysis of the user-generated content.
As an illustration, we present a distribution of geo-tagged Instagram posts in the center of St. Petersburg (Vasilevskiy Island), along with the distribution of the overall population density, calculated by the model.
Personal decision support mobile service for extreme situations // Procedia Computer Science. — 2014. — Vol. 29. — pp. 1646-1655.
Multiscale agent-based simulation in large city areas: emergency evacuation use case // Procedia Computer Science. — 2015. — Vol. 51. — pp. 2367-2376.
Optimization-based Calibration for Micro-level Agent-based Simulation of Pedestrian Behavior in Public Spaces // Procedia Computer Science. — 2015. — Vol. 66. — pp. 372-381.