cases-in-tn-by-age.R

library(tidyverse)
library(readxl)
d <- read_excel("Downloads/Public-Dataset-Age.xlsx")
d %>% 
    mutate(month = lubridate::month(DATE, label = TRUE)) %>% 
    group_by(month, AGE_RANGE) %>% 
    summarize(total_cases = sum(NEW_ARCASES)) %>% 
    filter(AGE_RANGE != "Pending") %>% 
    mutate(total_cases_prop = total_cases / sum(total_cases)) %>% 
    ggplot(aes(y = total_cases_prop, x = AGE_RANGE, fill = as.factor(month))) +
    geom_col(position = "dodge", alpha = 1) +
    scale_fill_viridis_d("") +
    coord_flip() +
    theme_minimal() +
    labs(x = "Age", y = "Proportion of Cases Per Month",
         title = "COVID-19 Cases in Tennessee by Age Range",
         subtitle = "Data: https://www.tn.gov/health/cedep/ncov/data/downloadable-datasets.html",
         caption = "created by Joshua Rosenberg")