The main purpose of this PhD thesis is to reach a better understanding of traffic operations behavior in urban networks. We want to furnish urban planners and traffic engineers with a general perspective of how street patterns affect traffic, which later, they can take into account when planning. This research is motivated by studying how planning of urban road space can be done in a more sustainable way, and by how innovative urban traffic management techniques can be more efficiently implemented. This first part of the dissertation is devoted to an in-depth study of street configurations on urban grid networks: two-way streets (TW), one-way streets (OW), and an intermediate solution, two-way streets with prohibited left turns (TWL). To date, no consensus has been reached on which street network configuration provides the optimal trade-off between these characteristics. On the one hand, strictly focusing on the movement of cars, two-way street networks provide a higher accessibility but also offer less capacity at intersections. On the other hand, one-way street networks provide intersections with higher capacities and travel speeds, but force drivers to travel longer routes on average. The analysis was carried out employing three different methods: analytical formulations, Static Traffic Assignment (STA), and Dynamic Traffic Assignment (DTA) based on microsimulation. The three methods present very different traffic assignment levels of complexity, computational times, and evidently research outcomes. Employing all of them makes this research robust and results more complete. TW networks provide the shortest distance traveled between two points and the highest route redundancy and flexibility. However, they present the lowest vehicle capacity at intersections which penalizes networks heavily. TWL networks offer the best trade-off between distance travelled and capacity at intersections. The main disadvantage is that they have a very limited route redundancy which translates into more heterogeneous spread of congestion. OW networks supply the highest capacity per movement at intersections although they also trigger the largest distance traveled. OW networks spread congestion more homogeneously than TWL networks because they have a higher route redundancy. The second part of this thesis deals with the removal of space in urban settings. Our aim is to understand and quantify how this removal affects drivers and the overall system. We also study abstract grid networks representing urban environments. First, we look at different strategies of link removal in the grid networks; later, we study the effects by only removing lanes. Same as in the first part of the dissertation, we employ two methods to model traffic: the STA and the microsimulation with a DTA module. Results clearly show that it is possible to remove certain amount of links from an urban grid without losing connectivity and worsening traffic conditions excessively. Evidently, traffic impacts depend on the removal strategy. Removing streets from the center creates the highest impacts because it affects the redundancy of routes, and hinders traffic flows from spreading more homogeneously. In that case, intersections with really high traffic loads will easily trigger congestion in the system. Removing streets from the perimeter, instead, allows the system to keep the high connectivity in the center that allows spreading traffic loads more evenly. Lane removal does not change the network connectivity but it makes traffic distribution more heterogeneous. Results show that central removal impacts the system more than peripheral removal in terms of capacity drop, congestion appearance and gridlock time. Congestion propagation depends on two factors: the number of one-lane streets in networks, and the origin of congestion. If congestion is originated in the perimeter, it spread over the network faster than if congestion started in the center. Although networks do not present clear congestion propagation patterns, central removal has a tendency to propagate congestion more from the center, whereas peripheral removal has a slight disposition to propagate congestion more from the perimeter. The third part of the dissertation dealt with the amount and location of fixed monitoring resources that a city should have to measure a reliable Macroscopic Fundamental Diagram (MFD) control scheme. This research is especially important for practitioners as we study a real case: the city of Zurich. Most cities do not have much traffic information a priori, so we propose different blind strategies to select only a certain amount of the links to be monitored to create an MFD. The results (obtained with simulation data) show that independently of the strategy used, a minimum of 25% of the links ensures a fairly accurate MFD. Selecting links randomly but with a certain weight towards central activity places (e.g. for the case of Zurich, the central train station) not only provides good results but also maintains a low variability, especially if at least 15% of the links are selected. If cities already have traffic counts or a simulation model where values of flows and densities can be estimated, the selection of links can be done in a very efficient way. Finally, like in the case of Zurich, some cities might already have some traffic counts for macroscopic control even if the coverage is not really high. Employing those same links or streets to create MFDs proved, for the case of Zurich, to be a very efficient strategy.