top of page

Research Interests

CAUSAL INFERENCE

potential outcome framework

Causality is a deeply debated philosophical concept that feels intuitive in everyday life, yet when we attempt to define it in mathematical or statistical terms, numerous fundamental challenges emerge.


The potential outcomes framework seeks to quantify the causal effect of a treatment or exposure on an outcome. To estimate these effects, we require causal assumptions and carefully specified statistical models.

​

Causal inference can be seen as a imputation problem of missing data.

domino_egg_causal (1).png

Bayesian Nonparametrics

infinite mixture models

Screenshot 2026-02-12 at 4.33.20 PM.png

  • Flexible distributions.

  • Uncertainty quantification.

  • Incorporate prior knowledge.

  • Handle missing data and imputation.

  • Variety of definitions: Gaussian process, Dirichlet process, Pitman–Yor process, BART, etc

Screenshot 2026-02-12 at 4.33.05 PM.png

Bayesian factor models

Screenshot 2026-02-12 at 4.44.46 PM.png

Factor models explain complex dependencies among correlated variables through a smaller number of latent factors, achieving crucial dimension reduction that makes high-dimensional problems tractable while improving interpretability and statistical efficiency.

Challenges: how do we define the distributions for the latent variable according with the research question?

Environmental health

  • How does air pollution affect human health?

  • What are the socio-economic characteristics that characterize the different vulnerabilities across the population?

  • Can air pollution regulation achieve the EPA's goal of environmental justice?

  • Is there an effect on social mobility if exposed to air pollution?

  • How do some events, such as wildfires, change the quality of the air?

Screenshot 2026-02-12 at 4.55.16 PM.png

Nutritional studies

Screenshot 2026-02-12 at 4.58.29 PM.png

Diet plays a crucial role in health outcomes, influencing the risk of chronic diseases such as cardiovascular disease, obesity, and diabetes. Understanding the causal effects of dietary interventions is essential for developing effective public health policies and personalized nutrition strategies.

​

One of the main challenges is that dietary intakes are typically collected as high-dimensional matrices of food or nutrients
consumption, with strong correlations among items, requiring principled dimensionality reduction.

bottom of page