A Brief Research Statement

My primary interests are in how large-scale economic, logistical, and institutional systems can be meaningfully measured using imperfect data. I am particularly drawn to problems where commonly used quantitative summaries obscure important system behavior, distributional effects, or institutional constraints. Rather than treating data as abstract inputs to models, my work emphasizes attention to data-generating processes, definitional heterogeneity, and the limits of inference in real-world settings.

Much of my applied experience comes from my time as a Data Research Fellow at the U.S. Department of Transportation, where I worked on a pilot program to release national freight statistics following the Ocean Shipping Reform Act of 2022. My work centered on shipping container chassis “dwell” data—records describing the timing of chassis movements within the intermodal freight system. Although the program aimed to improve transparency around port performance, the analysis quickly became a study in measurement constraints. I encountered substantial variation in how private-sector providers defined and reported dwells, persistent data quality issues, and confidentiality requirements that made publication impossible.

My contributions included building automated data validation and consolidation pipelines, exploring distributional anomalies in reported dwell times, and conducting disclosure analyses to determine when summary statistics could be released without risking provider identification. A key part of the work involved distinguishing meaningful signals from artifacts of reporting practices and communicating analytical limitations clearly to non-technical stakeholders. This experience shaped my interest in research questions at the intersection of statistical analysis, institutional design, and system-level interpretation, particularly in infrastructure, logistics, and public finance contexts.

In parallel with this applied policy work, I have explored measurement and interpretation challenges in a methodological setting through a project on plain-language classification using natural language processing. In that work, I constructed a labeled text corpus from scratch, explicitly addressing the ambiguity inherent in defining “plain language,” and trained baseline classifiers to distinguish plain from non-plain writing. While the modeling approach itself was intentionally simple, the project emphasized transparency about labeling subjectivity, model limitations, and evaluation uncertainty. I view this work less as an exercise in predictive optimization and more as an investigation into how qualitative concepts can be operationalized—and where such operationalizations break down.

Across these projects, a consistent theme in my work is an interest in measurement problems in complex systems: how institutional incentives shape data, how aggregation masks heterogeneity, and how analytical choices influence interpretation. Looking forward, I am particularly interested in questions such as how efficiency-optimized logistical or financial systems become fragile under scale and concentration, and where standard performance metrics in infrastructure or inequality research fail to capture distributional or temporal dynamics.

Methodologically, I approach these questions through applied statistical analysis of administrative, spatial, and textual data, combining exploratory modeling with close examination of data-generating processes. I do not yet have an extensive publication record, and I view a research-focused role as an opportunity to deepen my exposure to rigorous research workflows and collaborative problem formulation. My strengths lie in applied analysis, careful problem framing, and translating between technical methods and real-world systems.