My personal dissertation research focuses on 1) developing methodological frameworks to combine disparate spatiotemporal data with individual-level health data, and 2) using these linked datasets to estimate the effects of environmental exposures on health. To do this, I apply advanced quantitative methods in causal inference, statistical learning, geospatial analysis, and informatics to large-scale clinical, demographic, and geographic data.
In my role as a Graduate Research Assistant at UNC under Dr. Emily Pfaff, I leverage causal machine learning (ML) methods within a target trial framework to assess the comparative effectiveness of interventions as part of the National Clinical Cohort Collaborative (N3C) and the NIH RECOVER Initiative. Additionally, my work includes the development and validation of algorithmic approaches for automated clinical cohort identification and computable phenotype generation using graph analysis, natural language processing (NLP), and dimensionality reduction.
With Dr. Barbara Entwisle at the Carolina Population Center, I apply methods in computational demography to examine the potential of electronic health record data to inform research on population mobility dynamics.
As a Research Scientist at CrisisReady, I work with Drs. Satchit Balsari, Caroline Buckee, and Andrew Schroeder on building pipelines for crisis planning and response in the context of COVID-19, the Syrian War, and climate hazards in India. I am also interested in issues of human rights, ethics, and privacy, including differential privacy.
Examining whether prenatal lead exposure increases preterm birth risk, and whether micronutrient adequacy and community-level social vulnerability modify that association, using data from the ECHO Data Warehouse.
Comparing the predictive contribution of patient-reported social determinants captured in EHRs against Census-linked socioeconomic measures for glycemic control in a UNC Health type 2 diabetes cohort.
Electronic health records contain longitudinal patient address histories, a largely untapped, population-scale data source for studying residential mobility and spatial health inequity.
Linking spatially resolved sea-level rise exposure data to longitudinal residential histories from a TransUnion credit panel to estimate how coastal inundation risk drives household out-migration in North Carolina.
Using large-scale EHR data and target trial emulation to estimate the causal effects of well-water arsenic exposure on chronic disease outcomes, including bladder cancer and diabetes.
N3C aims to improve the efficiency and accessibility of analyses with COVID-19 clinical data, expand our ability to analyze and understand COVID, and demonstrate a novel approach for collaborative pandemic data sharing
The RECOVER Initiative is a patient-centered, integrated, adaptive research network
Data Readiness, Methods Readiness, and Translational Readiness, for equitable and transparent crisis planning and response.
A network of infectious disease epidemiologists at universities around the world working with technology companies to use aggregated mobility data to support the COVID-19 response
Evaluating the impact of the 2018 Kerala floods in partnership with the Govt. of Kerala, and the All India Disaster Mitigation Insitute
A rapid assessment household survey of Rohingya and host families in Cox's Bazar
A collaborative effort to optimize India's NCD management platform, aiming for streamlined workflows and improved outcomes across 18,000 centers.
A Harvard-wide research and policy collaborative focused on the development of a patient-centric, provider-friendly, API-enabled health exchange ecosystem.
Documenting the impact of the Syrian war on population morbidity and mortality, as part of the Lancet Commission on Syria
