Comparison to conversions to recovered (20 days gap, ggflags, gganimate)

cured_in_20days.R

library(tidyverse)
library(countrycode)  # to counvert country names to 2 letter codes
library(ggflags)      # for flags
library(gganimate)    # for video

# Getting data ------------------------------------------------------------

selected_countries <- c("Italy", "Spain", "France", "Germany", 
                        "Portugal", "United States", "Slovakia", "South Korea", "Japan", 
                        'Belgium', 'Sweden', 
                        'Switzerland', 'Denmark', 'Austria', 'Slovenia', 'Greece',
                        'Czechia', 'Romania', 'Ireland', 'Poland', 'Chile',
                        'Ukraine', 'Hungary', 'Finland', 'Albania', 'Bulgaria')

url_deaths <- "http://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv"
deaths <- read_csv(url_deaths) %>% filter(`Country/Region` %in% selected_countries, is.na(`Province/State`))

url_confirmed <- "http://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv"
confirmed <- read_csv(url_confirmed) %>% filter(`Country/Region` %in% selected_countries, is.na(`Province/State`))

url_recovered <- "http://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_recovered_global.csv"
recovered <- read_csv(url_recovered) %>% filter(`Country/Region` %in% selected_countries, is.na(`Province/State`))

# Comparison --------------------------------------------------------------

date1 <- "4/4/20"
active_cases = confirmed[[date1]] - deaths[[date1]] - recovered[[date1]]
active_cases

date2 <- "4/24/20"
new_recovered = recovered[[date2]] - recovered[[date1]]
new_recovered

cor(active_cases, new_recovered, method="spearman")

plot(log(active_cases), log(new_recovered))
abline(0,1, col="red", lty=2)

# data for plot
dt <- tibble(country = recovered$`Country/Region`, new_recovered = new_recovered, 
             active_cases = active_cases)
# adding 2 letter codes
dt$code <- countrycode(dt$country, origin = 'country.name', destination = 'iso2c')

# Plot --------------------------------------------------------------------

ggplot(dt, aes(x=active_cases, y=new_recovered, label=code)) +
  geom_label() +
  scale_x_log10(limits=c(1e2, 1.1e5)) +
  scale_y_log10(limits=c(1e2, 1.1e5)) +
  geom_abline(slope=1, intercept=0, color="red", lty=2) +
  labs(x = 'Active cases 20 days ago',
       y = 'Recovered during the last 20 days',
       title = 'Conversion of active cases to recovered in 20 days',
       subtitle = date2)
ggsave('cured_in_20days.png', width = 6, height = 6)

# Fun with flags ----------------------------------------------------------

ggplot(dt, aes(x=active_cases, y=new_recovered, country=tolower(code))) +
  geom_flag(size=7.5) +
  scale_x_log10(limits=c(1e2, 1.1e5)) +
  scale_y_log10(limits=c(1e2, 1.1e5)) +
  geom_abline(slope=1, intercept=0, color="red", lty=2) +
  labs(x = 'Active cases 20 days ago',
       y = 'Recovered during the last 20 days',
       subtitle = 'Conversion of active cases to recovered in 20 days',
       title = date2) +
  theme_bw()

# Video -------------------------------------------------------------------

dt_time = NULL

# data
for (i in 74:(ncol(deaths)-20)) {
  date1 <- names(deaths)[i]
  active_cases = confirmed[[date1]] - deaths[[date1]] - recovered[[date1]]
  
  date2 <- names(deaths)[i+20]
  new_recovered = recovered[[date2]] - recovered[[date1]]
  
  # data for plot
  dt <- data.frame(date = lubridate::mdy(date2), 
                   country = recovered$`Country/Region`, new_recovered = new_recovered, 
                   active_cases = active_cases)
  dt$code <- countrycode(dt$country, origin = 'country.name', destination = 'iso2c')
  dt_time <- rbind(dt_time, dt)
}  

# static plot
p = ggplot(dt_time, aes(x=active_cases, y=new_recovered, country=tolower(code))) +
  geom_flag(size=7.5) +
  scale_x_log10(limits=c(1e2, 1.1e5)) +
  scale_y_log10(limits=c(1e2, 1.1e5)) +
  geom_abline(slope=1, intercept=0, color="red", lty=2) +
  labs(x = 'Active cases 20 days ago',
       y = 'Recovered during the last 20 days',
       subtitle = 'Conversion of active cases to recovered (20 days gap)')
p

# video
p + transition_time(date) +
  labs(title = "Date: {frame_time}") +
  theme_bw()
anim_save(filename = "cured_in_20days.gif")