Spatial Forecasting of Ozone Levels

Spatial analysis makes some lovely graphs, doesn't it? The problem is that the datasets can be huge, which really forces you to ackowledge the practical side of model building. This project was a two-day take-home exam for an applied modeling course during my master's. The goal is simple: reconcile observed ozone levels from weather stations with predicted ozone levels from a CMAQ simulation using a Gaussian process, then make updated predictions for unobserved locations. There's a lot I would do to make this project better but given the short timeline and narrow page limit, I am proud of the overall quality and presentation.