Data engineering · Geospatial storytelling
Climate risk & housing
Explore how county housing markets relate to climate risk—and what changes around extreme weather events.
PythonDuckDBD3
Read the case study →Open interactive story ↗Projects
Questions first. Data, methods, and a story to follow.
These projects cover two sides of my work: using public data to explore real-world systems, and designing experiments to understand how machine-learning methods behave. Each case study links to the code and research behind it.
Data preparation · Exploratory analysis
Patterns of U.S. migration
Bring migration flows, county characteristics, and climate data together to investigate where people move.
PythonPublic dataCounty maps
Read the case study →Open research website ↗Machine learning · Experimental design
Neural recommender systems
Compare architectures, tuning methods, and optimizers under a shared training and resource budget.
PyTorchGraph learningOptimization
Read the case study →Explore the repository ↗From source data to something you can explore.
The interesting part rarely starts with a clean table. Across these projects, the work includes reconciling sources, checking geography and time, making comparisons fair, and explaining what the evidence can—and cannot—tell us.