Duiyi Dai (Claire)

Hi! I am an empirical economist and data scientist working at the intersection of economics, machine learning, large language models, and large-scale unstructured data.

I am currently an Empirical Research Analyst in Equity Research at UBS, where I use alternative datasets, quantitative methods, and data-driven research tools to study company and sector performance across European equities.

Previously, I was a Postdoctoral Researcher in Data Science and Informatics at the University of Oxford, working with Dr Charles Rahal on large-scale computational social science research.

I received my PhD in Economics from the University of Birmingham in 2025, supervised by Professor Robert Elliott and Dr Wanyu Chung.

My research develops novel measurement systems from large-scale text, image, and alternative datasets. I am particularly interested in economic uncertainty, AI and technological change, media economics, political economy, macroeconomics, and reproducible computational research. My work has been published or accepted in outlets including Patterns and the European Economic Review, and has been disseminated through CEPR and the Scottish Government.

You can find my CV here.

Publications & Forthcoming

“Visual Bias in the Brexit Referendum: A Quantitative Analysis of Newspaper Images” (Corresponding author, with Wanyu Chung and Robert Elliott). CEPR Discussion Paper 20524. Conditionally Accepted, European Economic Review.

“RobustiPy: An Efficient Next-Generation Multiversal Library with Model Selection, Averaging, Resampling, and Explainable AI” (with Daniel Valdenegro Ibarra, Jiani Yan, and Charles Rahal). Forthcoming, Patterns (Cell Press / Elsevier), 2026. arXiv; Code; Project site

Working Papers

“Measuring Brexit Uncertainty: A Machine Learning and Textual Analysis Approach” (Corresponding author, with Wanyu Chung and Robert Elliott). CEPR Discussion Paper 17410. Revised following referee reports from the Quarterly Journal of Economics; resubmission to the Economic Journal in preparation. Latest version (Here); Project webpage (Here); Policy brief (Here)

“The Macroeconomic Effects of Brexit Uncertainty” (Corresponding author, with Christoph Görtz). A continuation of our prior project, using a proxy-SVAR with a novel high-frequency instrument to identify Brexit uncertainty shocks. Draft

Work in Progress

“From Floods to Votes: The Impact of Natural Disasters on Electoral Outcomes through Social-Media Sentiment Analysis” (Corresponding author, with Wanyu Chung and Robert Elliott)

“The Visual Narrative of Climate Change: Assessing Its Effect on Capital Market Movements”

Software & Research Infrastructure

RobustiPy: an open-source Python library for multiverse and specification-curve analysis, with bootstrap inference, model selection and averaging, and SHAP-based explainability. Code; Project site

UK Parliamentary Debates & Speaker Biographies: a research-grade PostgreSQL database integrating historical Commons and Lords debates with structured speaker biographies, full-text search, and reproducible exports. Code

Teaching

Guest Lecturer, Aston University
2026

Invited Instructor, Textual Analysis and Machine Learning with Python (PhD), AIB UKI Conference
2024, Slides

Teaching Assistant, Global Economy (Undergraduates), University of Birmingham
2022, 2023, 2024

Teaching Assistant, International Trade Theory & Policy (Undergraduates), University of Birmingham
2021

Teaching Assistant, Econometrics with Stata (Undergraduates), University of Birmingham
2021, 2022