Home

I’m Luke Sonnet and I am the Lead Data Scientist at GrowthBook, the leading open source experimentation platform. My previous roles were working on central platform teams at Twitter (Experimentation Data Science) and Facebook (Demography and Survey Science). Before that I received my PhD in Political Science from UCLA.

Even before joining GrowthBook, I have enjoyed contributing to open source software and data sources, primarily in the domain of social science. At Facebook, I contributed to balance, a Python package for sample balancing, primarily via propensity score estimation. In graduate school, I was the primary author of estimatr, an R package of estimators commonly used by social scientists, and I contributed to the voteview.com project that estimates and hosts DW-NOMINATE ideal scores of US legislators. You can see more about the software and data I have worked on here.

My graduate student research focused on the behavior of legislators and voters. My research used largely experimental and quasi-experimental methods to study the effects of political communication, the role of political dynasties, and how social norms and female political participation. Most of my applied work is based in Pakistan.