Urban governance relies on complex data-driven operations and mechanisms to manage their scale. These systems can be designed for both efficiency and equity. How should it do so, in the presence of missing data, capacity constraints, and spatial heterogeneity -- while obeying transparency and operational constraints? We consider the design of two systems. (1) First, urban crowdsourcing, used to identify problems such as downed trees and power-lines. How should agencies respond to reports, when residents do not report problems at the same rates, and needs differ substantially spatially? (2) Second, congestion pricing. How should cities price the roadways, to ensure equitable access to transportation when ability to pay varies substantially? In each, we consider equity and efficiency tradeoffs operational decisions. We find that existing and naïve designs can reinforce existing socioeconomic disparities. However, by expanding the class of policies, substantial improvements on both metrics simultaneously are possible. Finally, I'll overview our work in translating the data science/machine learning insights to practice, to influence agency decision-making.
Copyright. All Rights Reserved.