covid_spain_tacker.R

# Ugly code for tacking Covid-19 in Spain

# load libraries
library(readr)
library(ggplot2)
library(tidyr)
library(dplyr)
library(grid)
library(gridExtra)
library(tabulizer)

# read data
url_confirmed <- "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_19-covid-Confirmed.csv"
corona <- readr::read_csv(url_confirmed)
url_deaths <- "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_19-covid-Deaths.csv"
corona_d <- readr::read_csv(url_deaths)

#corona <- filter(corona, `Province/State` == 'United Kingdom')
#corona_d <- filter(corona_d, `Province/State` == 'United Kingdom')


# drop unused columns
corona[c('Province/State', 'Lat', 'Long')] <- NULL
corona_d[c('Province/State', 'Lat', 'Long')] <- NULL

# from wide to long dataframe
corona <-
  pivot_longer(corona, cols = -c(`Country/Region`), names_to = 'date', values_to = 'cases') %>% 
  rename(country = `Country/Region`) %>%
  filter(country %in% c('Spain')) %>% 
  mutate(date = as.Date(date, format = '%m/%d/%y'))  


corona_d <-
  pivot_longer(corona_d, cols = -c(`Country/Region`), names_to = 'date', values_to = 'deaths') %>% 
  rename(country = `Country/Region`) %>%
  filter(country %in% c('Spain')) %>% 
  mutate(date = as.Date(date, format = '%m/%d/%y')) 


# set March 12 cases manually (non-updated data)
corona[corona$date == "2020-03-12",]$cases <- 3033
corona_d[corona_d$date == "2020-03-12",]$deaths <- 84

# join cases and deaths
corona <- 
  left_join(corona, corona_d, by = c('country','date')) %>% 
  mutate(fatality_rate = deaths/cases)

# create new cases variable
corona <-
  corona %>% 
  group_by(country) %>%
  mutate(daily_cases = cases - lag(cases)) %>% 
  ungroup()


# plots
p1 <-
  corona %>% 
  filter(date >= '2020-02-25') %>% 
  ggplot(aes(x = date, y = daily_cases)) +
  geom_point() + 
  geom_line() +
  labs(title = "New Daily Cases", 
       x = 'Date',
       y = 'Cases') +
  theme_minimal() +
  theme(plot.title = element_text(size=15))


p2 <-
  corona %>% 
  filter(date >= '2020-02-25') %>% 
  ggplot(aes(x = date, y = cases, label = cases)) +
  scale_y_log10() +
  geom_text(nudge_y = .15) +
  geom_point() +
  geom_line() +
  labs(title = "Coronavirus Cases", 
       x = 'Date', 
       y = 'Cases (log scale)') +
  theme_minimal() +
  theme(plot.title = element_text(size=15))


p3 <- 
  corona %>% 
  filter(date >= '2020-02-25') %>% 
  ggplot(aes(x = date, y = fatality_rate)) +
  scale_y_continuous(labels = scales::percent) +
  geom_point() + 
  geom_line() +
  labs(title = "Fatality rate", 
       x = 'Date', 
       y = 'Rate') +
  theme_minimal() +
  theme(plot.title = element_text(size=15))


p5 <- grid.arrange(p2,
             arrangeGrob(p1, p3, ncol = 2),
             nrow=2, 
             top = textGrob("Coronavirus in Spain", gp= gpar(fontsize=20, font = 2)))

ggsave(plot = p5, filename = "~/Descargas/covid_spain.png", width = 14, height = 10)



# Getting data for each CCAA 
# Thanks to 
# https://analisisydecision.es/seguimiento-del-coronavirus-en-espana-por-comunidad-autonoma-extraer-informacion-de-un-pdf-con-r/
pdf_file <- download.file(url = 'https://www.mscbs.gob.es/profesionales/saludPublica/ccayes/alertasActual/nCov-China/documentos/Actualizacion_43_COVID-19.pdf',
              destfile = tempfile(), method = 'curl', extra = "-k")
area <- locate_areas(pdf_file, pages = 2)


pdf_lista <- extract_tables(
  temp,
  output = "data.frame",
  pages = c(2),
  area = list(
    c(355.7635,  103.6804, 685.7704, 507.7376)
  ),
  guess = FALSE,
  encoding = "UTF-8"
)

datos_ccaa <- 
  data.frame(pdf_lista[1])[1:19, c('CCAA', 'Total.casos', 'Fallecidos')]

unlink(temp)


## Canarias
temp <- tempfile()
download.file(url = 'https://www3.gobiernodecanarias.org/sanidad/scs/content/dcb400c5-6504-11ea-9a8e-719d4b52bf6c/InformeCasosCOVID-19.pdf',
                          destfile = temp, method = 'curl', extra = "-k")

area <- locate_areas(temp, pages = 1)



pdf_lista <- extract_tables(
  temp,
  output = "data.frame",
  pages = c(1),
  area = list(
    c(393.4206, 124.5549, 536.8005, 394.7037 )
  ),
  guess = FALSE,
  encoding = "UTF-8"
)


df <- data.frame(pdf_lista[[1]])
df <-
  df %>% 
  rename(Edad = Grupo.de.Edad) %>% 
  select(Edad, Mujer, Hombre)

df$Edad <- factor(df$Edad, levels = c("10-19 años", "20-29 años", "30-39 años", "40-49 años", "50-59 años", "60-69 años","70-79 años" ,"80-89 años", ">=90 años" ), ordered = TRUE)

p <-
  df %>% 
  pivot_longer(cols = c('Mujer', 'Hombre'), names_to = 'Genero', values_to = 'Casos') %>% 
  ggplot(aes(x = Edad, y = Casos, fill = Genero)) +
  geom_bar(stat="identity", position=position_dodge())+
  theme_minimal() +
  scale_fill_manual(values=c('#999999','#E69F00')) +
  geom_text(aes(label = Casos), vjust=-1, color="black",
            position = position_dodge(0.9), size=3.5) +
  labs(title = 'COVID-19 en Canarias', 
       subtitle = 'Número de casos detectados por grupo de edad y género',
       fill = 'Género', 
       caption = "Fuente: Consejería de Sanidad, Gobierno de Canarias\nActualización: 16/03/2020") +
  theme(plot.title = element_text(size=20), 
        plot.subtitle = element_text(size = 17),
        axis.text =  element_text(size=14),
        axis.title = element_text(size=14),
        legend.title = element_text(size=14),
        legend.text = element_text(size=13))
  

ggsave(plot = p, filename = "~/Descargas/covid_canarias.png", width = 14, height = 10)