My research sits at the intersection of quantitative marketing, information systems and econometrics: I study how digital platforms such as music streaming and social media reshape consumer behavior and markets, and build the methods to measure it. My background as a software and information engineer enables me to work with large-scale data under limited computational resources, a quiet but essential foundation of most of my projects. Methodologically, I combine representation learning, causal inference, and natural language processing. My work on emoji-based emotion detection is forthcoming at the Journal of Marketing.
I have held research visits at NYU Stern, Bocconi, UNSW, and Imperial College London, and review for the International Journal of Research in Marketing.
From papers to tools
Research should be reusable, not just readable. Several of our projects come with open tools and packages you can run on your own data.
- NADE Explorer: type a sentence and see the emotion it expresses, right in the browser. From our Journal of Marketing paper; also available as Python and R packages.
- DiD Estimand Lab: an interactive companion to our difference-in-differences (DiD) paper. Move the sliders to see how the estimands and estimators shift under heavy-tailed outcomes, with a replication package in Python, R, and Julia. See the paper.
In the press
Coverage of the emoji-based emotion-detection work, across trade, science, and broadcast media:
- American Marketing Association: a new tool that reads emotion from emojis (2026)
- Phys.org: science-press coverage of the method (2026)
- The Last Show with David Cooper: featured discussing the work on “Linguistics from a Linguist” (2026)