County sheriff elections & incarceration

Sheriffs are among the most powerful and least studied elected officials in America. Before this project there was no national dataset of county sheriff elections. With a research team in Stanford Medicine’s Department of Epidemiology and Population Health, I helped build the first one.

The data work was messy in the way civic records always are: thousands of counties, wildly inconsistent formats, and every design decision a tradeoff between coverage and confidence. On top of the dataset I built inferential models to surface links between sheriff demographics and local incarceration rates, and produced exploratory visualizations documenting racial disparities, meant to inform diversity recruiting and policy conversations.

It was my first taste of research where the stakes are civic rather than commercial, and where being careful about what the data can’t say matters as much as what it can.