# scrapes and plots some key covid metrics for florida

florida-covid-19

library(dplyr)
library(rvest)
library(stringr)
library(ggplot2)
library(tidyr)

STATE_OF_INTEREST <- "Florida"

STATE_OF_INTEREST <- str_to_lower(STATE_OF_INTEREST)

url <- glue::glue("https://covidtracking.com/data/state/{STATE_OF_INTEREST}#historical")

fla_html <- read_html(url)

fla_dat <- fla_html %>% html_table() %>% `[[`(2)

fla_dat <- 
  fla_dat %>% 
  rename_all(tolower) %>%
  mutate(
    date = as.Date(date, "%a %B %d %Y")
  ) %>%
  mutate_at(
    vars(-c(date, starts_with("screenshot"))), .funs = function(x) {as.numeric(str_replace_all(x, "[[:punct:]]", ""))}
  ) %>%
  arrange(date) %>%
  mutate(
    new_positive = c(NA, diff(cases)),
    new_hospitalized = c(NA, diff(hospitalized)), 
    new_deaths = c(NA, diff(deaths))
  )

fla_dat %>% 
  mutate(
    frac_positive = new_positive / `new tests`
    ) %>% 
  filter(`new tests` > 100) %>% 
  ggplot(aes(x = date, y = frac_positive)) + geom_point() + geom_line()

fla_dat %>% ggplot(aes(x = date, y = new_deaths)) + geom_point() + geom_line()
fla_dat %>% ggplot(aes(x = date, y = new_positive)) + geom_point() + geom_line()

fla_long <- 
  fla_dat %>% 
  dplyr::select(date, new_deaths, new_positive, `new tests`, new_hospitalized) %>% 
  pivot_longer(cols = c(new_deaths, new_positive, `new tests`, new_hospitalized), names_to = "metric") %>%
  group_by(metric) %>%
  mutate(
    scaled_value = (value - mean(value, na.rm = T)) / sd(value, na.rm = T)
  )

fla_long %>% 
  ggplot(aes(x = date, y = scaled_value, color = metric)) + 
  geom_point(alpha = .3) + 
  geom_line(alpha = .3) + geom_ma(n = 7, linetype = "solid", size = 1.3) + 
  ggtitle(glue::glue("{STATE_OF_INTEREST} covid stats over time"))