covid-canada.R

covid-canada.R

library(tidyverse)
library(gghighlight)
library(ggrepel)
library(scales)

data <- read_csv("https://github.com/CSSEGISandData/COVID-19/raw/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv")

data_long <- pivot_longer(data, `1/22/20`:last_col(), names_to = "date", values_to = "cases") %>%
  mutate(date = as.Date(date, format = "%m/%d/%y")) %>%
  filter(!is.na(cases),
         (`Province/State` != "Grand Princess" | is.na(`Province/State`)),
         `Country/Region` != "Cruise Ship")


data_by_country <- group_by(data_long, country = `Country/Region`, date) %>%
  summarise(cases = sum(cases, na.rm = TRUE))

highlight_countries <- c("Canada", "US", "Italy", "United Kingdom", "Spain", "Australia")

(plot_countries_by_date <- filter(data_by_country, country != "China") %>%
  ggplot(aes(x = date, y = cases, colour = country)) +
  scale_x_date(date_breaks = "1 week", date_labels = "%b %d") +
  scale_y_log10(expand = c(0,0.5),
                labels = label_comma(accuracy = 1)) +
  geom_line(size = 1.5) +
  gghighlight(country %in% highlight_countries,
              use_direct_label = FALSE,
              unhighlighted_params = list(size = 0.5)) +
  geom_text_repel(aes(label = country),
                  data = filter(data_by_country,
                                country %in% highlight_countries,
                                date == max(date)),
                  nudge_x = 1, direction = "y",
                  segment.size = 0,
                  xlim = c(max(data_by_country$date),
                           max(data_by_country$date) + 14)) +
  scale_color_brewer(palette = 2, type = "qual", guide = "none") +
  coord_cartesian(clip = "off") +
  theme_minimal() +
  theme(plot.margin = unit(c(0.5,3,0.5,0.5), "cm")) +
  labs(x = "Date",
       y = "Total number of cases (log scale)",
       title = "COVID-19 Confirmed Cases"))

n_cases <- 100

data_by_country_n_days <- data_by_country %>%
  filter(cases >= n_cases) %>%
  group_by(country) %>%
  arrange(country, date) %>%
  mutate(days_since_nth_case = as.numeric(date - min(date, na.rm = TRUE))) %>%
  ungroup()


(plot_100_cases <- filter(data_by_country_n_days, country != "China") %>%
  ggplot(aes(x = days_since_nth_case, y = cases, colour = country)) +
  scale_y_log10() +
  geom_line(size = 1.5) +
  gghighlight(country %in% highlight_countries,
              use_direct_label = FALSE,
              unhighlighted_params = list(size = 0.5)) +
  geom_text_repel(aes(label = country),
                  data = filter(data_by_country_n_days,
                                country %in% highlight_countries,
                                date == max(date)),
                  nudge_x = 1, direction = "both",
                  segment.size = 0) +
  scale_color_brewer(palette = 2, type = "qual", guide = "none") +
  theme_minimal() +
  labs(x = glue::glue("Days since reaching {n_cases} confirmed cases"),
       y = "Total number of cases (log scale)",
       title = glue::glue("COVID-19 confirmed cases since reaching {n_cases} confirmed cases")))

data_by_country_1st_case <- data_by_country %>%
  filter(cases >= 1) %>%
  group_by(country) %>%
  arrange(country, date) %>%
  mutate(days_since_nth_case = as.numeric(date - min(date, na.rm = TRUE))) %>%
  ungroup()


(plot_1st_cases <- filter(data_by_country_1st_case, country != "China") %>%
  ggplot(aes(x = days_since_nth_case, y = cases, colour = country)) +
  scale_y_log10() +
  geom_line(size = 1.5) +
  gghighlight(country %in% highlight_countries,
              use_direct_label = FALSE,
              unhighlighted_params = list(size = 0.5)) +
  geom_text_repel(aes(label = country),
                  data = filter(data_by_country_1st_case,
                                country %in% highlight_countries,
                                date == max(date)),
                  nudge_x = 2, direction = "both",
                  segment.size = 0) +
  scale_color_brewer(palette = 2, type = "qual", guide = "none") +
  theme_minimal() +
  labs(x = glue::glue("Days since first confirmed case"),
       y = "Total number of cases (log scale)",
       title = "COVID-19 confirmed cases since first confirmed case"))

canada <- rename(data_long,
                 country = `Country/Region`,
                 province = `Province/State`) %>%
  filter(country == "Canada")

(plot_canada <- ggplot(canada, aes(x = date, y = cases, colour = province)) +
  geom_line(size = 1) +
  scale_x_date(date_breaks = "1 week", date_labels = "%b %d") +
  scale_y_continuous(expand = c(0,1)) +
  scale_color_brewer(palette = 3, type = "qual", guide = FALSE) +
  coord_cartesian(clip = "off") +
  geom_text_repel(aes(label = province),
                  segment.size = 0,
                   data = filter(canada, date == max(date)),
                   nudge_x = 1, direction = "y",
                  xlim = c(max(canada$date), max(canada$date) + 14)
                  ) +
  theme_minimal() +
  theme(plot.margin = unit(c(0.5,3,0.5,0.5), "cm")) +
  labs(x = "Date",
       y = "Total number of cases",
       title = "COVID-19 Confirmed Cases in Canada"))

max_date <- max(data_long$date)

out_dir <- "~/Desktop/covid-plots"
dir.create(out_dir, showWarnings = FALSE)
ggsave(glue::glue("{out_dir}/countries_by_date-{max_date}.png"), plot_countries_by_date)
ggsave(glue::glue("{out_dir}/100_cases-{max_date}.png"), plot_100_cases)
ggsave(glue::glue("{out_dir}/1st_cases-{max_date}.png"), plot_1st_cases)
ggsave(glue::glue("{out_dir}/canada-{max_date}.png"), plot_canada)