Urban Traffic Forecasting, Urban Data Platforms and Urban Foundations Models

In the context of urban environments, accurate traffic forecasting is of significant importance to a multitude of stakeholders. This study aims to demonstrate the potential of sensors installed within the traffic network to facilitate precise forecasting. The work addresses various aspects, including sensor types, sensor positioning, data and tool pipelines, and the utilization of graph neural networks (GNN) for traffic forecasting. Additionally, it offers a perspective on applying these networks in urban foundation models (UFM) for forecasting purposes. This work illustrates the required sensors, tasks, and methods for traffic forecasting based on the city of Osnabrück in Germany as a case study.