Project II: Analysis of Rates of Vaccination of COVID-19

Project II: Abstract

COVID-19 has greatly effected the livelihood of many in the United States. It has become more important than ever to recognize the effect of vaccines on the public. Of interest is to correctly estimate the individuals who are fully vaccinated. It is important to keep in mind how individuals are vaccinated not only by state but also by time zone. In this exercise, we are interested in constructing two models to understand the proportion of individuals who are fully vaccinated and the vaccination rate per state and time zone. Model 1 is seen as the simpler model, and model 2 is seen as the more complex model. After analysis, and model comparison, we conclude that model 2 is the more robust model providing a better analysis, and being favored by Bayes factor, DIC, and PPLC.

Exploratory Data Analysis II

In this project we want to visualize how many people are fully vaccinated across the main time zones in the states and we do so with a color coded scatterplot similar as the first project.

We would expect to the see the most populated states for their respective time zone have the most vaccine counts and that is exactly what we see. We note that New York and Florida are the most vaccinated in EST, for PST we have California, for CT there is Texas, and MT there is Arizona.

Statistical Modeling and Analysis II

We proceed with comparing both models by first computing the marginal likelihoods of both models to compute the Bayes factor, as well as using DIC and PPLC tools. Trace plots of the Markov chains were used to assess that convergence was reached. Model 2 being the more complex model by creating a hierarchical modeling structure appears to be more robust.