My GitHub has code, written summaries of projects I’ve worked on, and slides from some presentations I’ve given. Some highlights:
- I developed an R package alongside my advisor Jacob Bien implementing cluster stability selection. Cluster stability selection is a feature selection method we developed that mitigates the issues with stability selection when important features are highly correlated.
- My team tied for second place in the 2020 COVID-19 Computational Challenge, hosted by the City of Los Angeles and RMDS. My code and our written report are available on my GitHub. We developed a model to forecast new COVID infections every day in 78 neighborhoods across Los Angeles. We used the SIR model alongside the best COVID research available at the time to estimate daily infections from positive test results. Then we used a Poisson autoregressive model to forecast new infections. Finally, we used synthetic controls to forecast infections for neighborhoods with missing data.
- I have slides and written reports available for some topics I have studied in detail but not yet conducted research on, including multitask learning and algorithmic fairness.
- I also have code from less recent projects I’ve worked on, including my contribution to the Fragile Families Challenge (see the research paper which I co-authored), and my team’s entry in the Orange Country R Users Group Hackathon 2019, which won “Best Model.”