We present the Decision Support Tool (DST) prototype that is being developed in partnership with two regional companies: Energies Demain and Logiroad, which aims to predict how freight transportation flows will be impacted by major projects in city centers, such as the implementation of new pedestrian areas, limited traffic zones, or new logistics facilities. The project is still in its initial phases, so based on a case study using data from the city of Nantes, we focus on presenting the two main components of the future DST: the estimation of freight delivery flows in a city center, and the validation of this estimation with sensors. For the flow estimation we solve a series of Vehicle Routing Problems with Time-Windows using a Large Neighborhood Search metaheuristic. The solutions are then plotted in the same city network, and the flow is given by the number of routes using the same network link, at a certain moment (or time interval). With respect to the validation, sensors based on image recognition will be installed in the city, so the real number of delivery vehicles can be counted. In this regard, we have to deal with the Sensor Location Problem. We conclude the presentation by the roadmap of the research project, which integrates data science, operational research and image recognition.