I investigated motorcycle fatalities and intercorrelations between various forms of unsafe operator behavior. This project included a lot of data exploration and visualization, as well as some random forest classifiers. See my writeup of the project on my blog here, as well as the code on github.

I used Gaussian process regression and a few other tools to create a model of concentrations in a large chemical spill underground! Follow this link to see my writeup of this project on this blog. Code and a powerpoint presentation can be found at my github, here.

I investigated whether you could predict selenium concentrations in seafood using only the information on a USDA nutrition label, in an easily-interpretable way. See my writeup on this blog here, as well as the code and a powerpoint presentation on github.