Part 1 - Preliminary Data Wrangling The below code chunk reads in the data, removes Alaska and Hawaii (we are only working with the contiguous United States) and saves it as county_data . It also reports the proportion of counties with 0 deaths and 0 cases. county_data % filter (!State %in% c ("AK", "HI") ) county_data >% summarize (Proportion_county_no_cases = mean (Total_cases==0), Proportion_county_no_deaths = mean (Total_deaths==0) ) %>%% kable () Proportion_county_no_cases Proportion_county_no_deaths 0.0006435 0.0814028 You should note that a non-negligible proportion of counties in the United States have reported no Coronavirus related deaths, whereas few counties have reported no cases at all. This causes our response variable for the day, death rates in counties, to be zero inflated - modeling this sort of data is incredible important in science and society but is outside the scope of this course. In module 6 we handled the few zero cases by arbitrarily adding a constant (i.e., 1) to all the responses. Technically we could do that here, but then we would have 1-inflated data - it would not really fix the underlying "inflation" problem. In the below code chunk, filter the data such that only counties with at least 1 death are retain. Note, by only working with a subset of the data, we now have a different target population than the original sample. Describe what the target population for the inference and analysis using this data. ANSWER HERE