Selected projects in statistics and machine learning

---

Using Bayesian spline regression to estimate effects of lactation on maternal immune function

While the impact of breastfeeding behavior on infant health has been well-studied, less attention has been paid to the effects of lactational practices on maternal health. For this peer-reviewed publication, I utilized data I collected as co-Principal Investigator of the Seattle Postpartum Health Study and Bayesian spline regression models (using the brms package in R) to show that breastfeeding is associated with non-linear effects on diurnal changes in immune function.

Publication prepared using R Markdown, code published on Zenodo and GitHub.

---

Using GAMs to compare sex differences in immune function in two distinct populations

Research from industrialized societies like the USA show that women are less prone to infectious disease but suffer disproprotionately from autoimmune disease. Hormones (e.g., testosterone, estrogen) are large drivers of these sex differences in immune function. Because ovarian hormones (e.g., estraidol, progesterone) are particularly sensitive to ecological conditions (e.g., nutrition), sexual dimorphism in immune function might be less severe in non-industrialized populations. For this peer-reviewed paper, I used preexisting data from NHANES and the Tsimane Health and Life History Project, propensity score matching via KNN, and generalized additive models (GAMs) using the mcgv package in R to show that the degree of sex difference in immune function is heavily influenced by environment.

Publication prepared using R Markdown, code published on GitHub.

---

ML in Python, MIMIC III, IV, and eICU data

Project description in two sentences.

Read report on Quatro