Modern societal systems are increasingly complex and present challenges in (1) understanding their impact through the vast amount of data collected, and (2) designing policies that promote efficient and equitable usage. In this dissertation, we describe methods and applications that tackle these challenges in two real-world systems: the urban crowdsourcing system, and the urban library system. We demonstrate how to leverage robust theoretical modeling and empirical analyses together, to uncover insights from data, and inform policy design. Part I studies the urban resident crowdsourcing system, which city governments are increasingly relying on to identify problems such as downed trees and power lines. We first develop a method to identify heterogeneous usage by residents, without using external ground-truth data. Our main insight is that duplicate reports about the same incident can be leveraged to disambiguate whether an incident has occurred with its reporting rate once it has occurred, and transform the problem into a Poisson process rate estimation task. We apply our method to over 100,000 resident reports from the New York City 311 system and to over 900,000 reports from the Chicago 311 system, finding that there are substantial spatial disparities in how quickly incidents are reported, even after controlling for incident characteristics, which further correspond to socioeconomic characteristics. We then consider the allocation of government resources to these resident reports, which needs to be efficient and equitable -- though these desiderata may conflict. In particular, we consider the design of Service Level Agreements (SLA): promises that reports will be responded to within a certain time. The city has two decision levers: how to allocate response budgets to different neighborhoods, and how to schedule responses to individual incidents. We theoretically analyze a stylized model, finding that the equity-efficiency trade-off may be substantial in realistic settings, and centralizing response efforts leads to relatively small improvements. We then develop a simulation-optimization framework to empirically optimize policies in NYC, and find that (a) status quo inspections are highly inefficient and inequitable compared to optimal ones, and (b) in practice, despite large trade-off, the empirical "price of equity'' is small -- large improvements in equity is possible with relatively small decreases in efficiency. Part II studies the urban library system. We first propose a Bayesian framework to characterize book checkout behavior across multiple branches of a library system, learning heterogeneous book popularity, overall branch demand, and usage of the online hold system, while controlling for book availability. In collaboration with the New York Public Library, we apply our framework to granular data consisting of over 400,000 checkouts during 2022. We show that disparities are largely driven by disparate use of the online holds system, which allows library patrons to receive books from any other branch through an online portal. This system thus leads to a large outflow of popular books from branches in lower-income neighborhoods to those in high-income ones. We then tackle the optimization of the holds system, seeking to retain its benefits in increasing usage, while balancing browser experience in lower-income branches. The library has two levers: where a book should come from when a hold request is placed, and how many book copies at each branch should be available through the holds system versus reserved for browsers. We first show that the problem of maximizing usage can be viewed through the lens of revenue management, for which near-optimal fulfillment policies exist. We then develop a simulation framework that further optimizes for browser experience, through book reservations. Empirically, we find that though a substantial trade-off exists between these two desiderata, a balanced policy can improve browser experience over the historical policy without significantly sacrificing usage. Because browser usage is more prevalent among branches in low-income areas, this policy further increases system-wide equity: notably, for branches in the 25% lowest-income neighborhoods, it improves both usage and browser experience by about 15%. Throughout this dissertation, we combine statistical analyses into the structures in the data-generating process with fine-grained data to understand the two systems, and combine stylized operational models with computational methods to optimize robust policy designs for them. We hope the methods developed and their applications demonstrate the value of these approaches, and provide a reference for future work that designs more efficient and equitable societal systems.