Mobile Traffic Time Series: Urban Region Representations and Synthetic Generation

The aim of this work is to build a methodology for representing urban regions using service-specific mobile traffic data from the Netmob dataset. Despite a rich literature on the topic, this kind of data has never been used to this purpose and it can therefore provide valuable insights into the dynamics and patterns of urban activity that were previously unexplored. Key steps include embedding time series data through autoencoding and utilizing different aggregation techniques to build geographically contextualized representations. As a further step we explore possible solutions to synthetically generate this kind of data in new urban areas.