Projects

Things I've built

Coursework and independent projects, grouped into classical statistical methods and machine learning and Bayesian work. Each card links to the full write-up.

Classical Statistics
Medical

COVID-19 Diagnosis from Lab Tests

Logistic regression on 608 hospital patients to find laboratory markers and demographics that predict COVID-19 infection when RT-PCR testing is unavailable.

View report →
Medical

Smoking, Age & Infant Mortality

A clinical contingency-table study using tests for proportions to relate a mother's age and smoking habits to gestational age and infant mortality.

View report →

Count Regression with GLMs

Quasi-binomial and negative-binomial models for overdispersed count data, applied to glove-use and school-absence datasets.

View report →

Popcorn Interest Time Series

SARIMA modeling of 2004 to 2021 Google Trends data, using transformation and seasonal differencing to forecast demand.

View report →
Machine Learning & Bayesian Methods

Bayesian Regularized Neural Networks

Replicating and extending Bayesian regularized neural networks for stock-market forecasting, testing where they outperform plain feed-forward nets.

View report →

Naive Bees: CNN Image Classification

A convolutional neural network in Keras and TensorFlow that classifies honey bees versus bumblebees from 1,654 field images.

View poster →
← Experience Misc →