dashboard-peru-covid-19.Rmd

--- 
title: "COVID-19 en Perú (v0.9.5) #YoMeQuedoEnCasa "
#author: "[Jesus M. Castagnetto]"
output: 
  flexdashboard::flex_dashboard:
    css: style.css
    orientation: columns
    vertical_layout: fill 
    theme: united
    navbar:
      - {title: "", icon: "fab fa-twitter", href: "https://twitter.com/share?text=COVID-19+en+Perú&url=https://castagnetto.site/peru/dashboard-peru-covid-19.html", target: "_blank"}
      - {title: "", icon: "fab fa-linkedin", href: "https://www.linkedin.com/shareArticle?mini=true&url=https://castagnetto.site/peru/dashboard-peru-covid-19.html&title=COVID-19+en+Perú", target: "_blank"}
      - {title: "Código fuente y Datos",  icon: "fab fa-github", href: "https://github.com/jmcastagnetto/covid-19-peru-data", target: "_blank"}
    includes:
      in_header: _ga.html
---

```{r setup, include=FALSE}
library(flexdashboard)
library(tidyverse)
library(patchwork)
library(DT)
library(echarts4r)
library(echarts4r.maps)
library(showtext)
#library(plotly)
#library(ggiraph)
#library(ggiraphExtra)
library(lubridate)
library(latex2exp)
library(runner)

font_add_google("Inconsolata", "inconsolata")
font_add_google("Lato", "lato")

showtext_auto()

knitr::opts_chunk$set(
  message = FALSE,
  warning = FALSE,
  echo = FALSE
)

last_updated <- paste0(
  "Actualizado el ",
  lubridate::now(tzone = "UTC"), 
  " UTC"
)

jmc <- "@jmcastagnetto, Jesús M. Castagnetto"
minsa <- "Fuente: @Minsa_Peru (https://twitter.com/Minsa_Peru) y comunicados del MINSA"
no_deaths_minsa_20200412 <- ""
#no_deaths_minsa_20200412 <- "Desde el 2020-04-12, MINSA ya no publica el número de fallecimientos por región"

covid_pe <- read_csv(
  "datos/covid-19-peru-data.csv",
  col_types = cols(
    country = col_character(),
    iso3c = col_character(),
    region = col_character(),
    date = col_date(format = ""),
    confirmed = col_double(),
    deaths = col_double(),
    recovered = col_double(),
    total_tests = col_double(),
    negative_tests = col_double(),
    pcr_test_positive = col_double(),
    serological_test_positive = col_double(),
    ag_test_positive = col_double(),
    pcr_serological_test_positive = col_double()
  )
) %>% 
  filter(is.na(region)) %>% 
  select(-region)

covid_regions_pe <- read_csv(
  "datos/covid-19-peru-data-augmented.csv",
  col_types = cols(
    country = col_character(),
    iso3c = col_character(),
    region = col_character(),
    region_orig = col_character(),
    ubigeo = col_character(),
    iso_3166_2_code = col_character(),
    date = col_date(format = ""),
    confirmed = col_double(),
    deaths = col_double(),
    recovered = col_double(),
    total_tests = col_logical(),
    negative_tests = col_double(),
    pcr_test_positive = col_double(),
    serological_test_positive = col_double(),
    ag_test_positive = col_double(),
    pcr_serological_test_positive = col_double(),
    zone = col_character(),
    pop2020 = col_double(),
    surface = col_double(),
    capital = col_character(),
    altitude = col_double(),
    lat = col_double(),
    lon = col_double()
  )
  # 
  # 
  # col_types = cols(
  #   country = col_character(),
  #   iso3c = col_character(),
  #   region = col_character(),
  #   region_orig = col_character(),
  #   cod_dep_inei = col_character(),
  #   cod_dep_reniec = col_character(),
  #   iso_3166_2_code = col_character(),
  #   date = col_date(format = ""),
  #   confirmed = col_double(),
  #   deaths = col_double(),
  #   recovered = col_double(),
  #   total_tests = col_double(),
  #   negative_tests = col_double(),
  #   pcr_test_positive = col_double(),
  #   serological_test_positive = col_double(),
  #   pcr_serological_test_positive = col_double(),
  #   pop2020 = col_double()
  # )
) %>%
  filter(!is.na(region)) %>% 
  group_by(region, ubigeo,
           iso_3166_2_code, date, pop2020) %>% 
  summarise(
    confirmed = sum(confirmed, na.rm = TRUE),
    deaths = sum(deaths, na.rm = TRUE),
    total_tests = sum(total_tests, na.rm = TRUE),
    negative_tests = sum(negative_tests, na.rm = TRUE),
    pcr_test_positive = sum(pcr_test_positive, na.rm = TRUE),
    serological_test_positive = sum(serological_test_positive, na.rm = TRUE),
    ag_test_positive = sum(ag_test_positive, na.rm = TRUE),
    pcr_serological_test_positive = sum(pcr_serological_test_positive, na.rm = TRUE)
  ) %>%
  distinct() %>% 
  mutate(
    confirmed_per_million = round(confirmed * 1e6 / pop2020, 1)
  )

cur_theme <- theme(
  plot.margin = unit(rep(1, 4), "cm"),
  plot.title = element_text(family = "lato"),
  plot.subtitle = element_text(family = "lato"),
  plot.caption = element_text(family = "inconsolata"),
  legend.position = "none"
)
```

# Situación general {data-orientation=rows}


```{r}
sit_pe <- covid_pe %>% 
  slice_tail(n = 3) %>% 
  select(-country, -iso3c) %>% 
  arrange(date) %>% 
  mutate_at(
    2:6,
    format,
	big.mark = ",",
	nsmall = 0
  )

```


```{r}
## General info
### Periodo de cuarentena {data-width="70%"}
susp_classes_first_day <- ymd("2020-03-12", tz = "America/Lima")
susp_classes_last_day <- ymd("2020-06-30", tz = "America/Lima")
decree_day <- ymd("2020-03-15", tz = "America/Lima")
first_day <- ymd("2020-03-16", tz = "America/Lima")
last_day <- ymd("2020-06-30", tz = "America/Lima")
today <- now(tzone = "America/Lima")

cuarentena_df <- data.frame(
  id = 1:2,
  title = c(
    paste("Suspensión de clases: del 2020-03-12 al 2020-06-30"), 
    paste("Cuarentena: del 2020-03-16 al 2020-06-30")
  ),
  content = c("Suspensión de clases", "Cuarentena COVID-19, Perú"),
  start = c(susp_classes_first_day, first_day),
  end = c(susp_classes_last_day, last_day),
  group = c("susp", "quar"),
  style = c("background-color:yellow", "background-color:cyan")
)
# timevis::timevis(cuarentena_df, fit = TRUE, showZoom = FALSE)

# last 4 records, yesterday and today
beds_icu <- read_csv("datos/covid-19-peru-camas-uci.csv") %>% 
  slice_tail(n = 6) %>% 
  select(fecha, estado, total) %>% 
  pivot_wider(
    names_from = estado,
    values_from = total
  ) %>% 
  janitor::clean_names() %>%
  mutate_at(
    vars(en_uso), #disponible),
    scales::comma
  )
```


```{r eval = FALSE}
### {data-width="30%"}
# **Días en cuarentena oficial**
ahora <- (interval(first_day, today) %/% days()) + 1
final <- (interval(first_day, last_day) %/% days()) + 1
# gauge(ahora, min = 0, max = final, symbol = "d", 
#       gaugeSectors(colors = "darkgrey"))
```


```{r}
# **Días en suspensión de clases**
ahora <- (interval(susp_classes_first_day, today) %/% days()) + 1
final <- (interval(susp_classes_first_day, susp_classes_last_day) %/% days()) + 1
# gauge(ahora, min = 0, max = final, symbol = "d", 
#      gaugeSectors(colors = "blue"))
```

## Anteayer

### Anteayer

```{r}
valueBox(sit_pe[1,]$date, color = "#ffffff")
```

### Casos confirmados

```{r}
valueBox(sit_pe[1,]$confirmed, icon = "fa-user-plus", color = "info")
```


### Fallecimientos

```{r}
valueBox(sit_pe[1,]$deaths, icon = "fa-cross", color = "#ff0000")
```

### Casos Recuperados

```{r}
valueBox(sit_pe[1,]$recovered, icon = "fa-user-check", color = "#ffa500")
```

### Casos Negativos

```{r}
valueBox(sit_pe[1,]$negative_tests, icon = "fa-user-minus", color = "success")
```

### Camas UCI en uso
```{r}
valueBox(beds_icu[1,]$en_uso, icon = "fa-procedures", color = "#e57373")
```

### Camas UCI disponibles
```{r}
valueBox(beds_icu[1,]$disponible, icon = "fa-bed", color = "#4dd0e1")
```


## Yesterday

### Ayer
```{r}
valueBox(sit_pe[2,]$date, color = "#ffffff")
```


### Casos confirmados
```{r}
valueBox(sit_pe[2,]$confirmed, icon = "fa-user-plus", color = "info")
```

### Fallecimientos

```{r}
valueBox(sit_pe[2,]$deaths, icon = "fa-cross", color = "#ff0000")
```

### Casos Recuperados
```{r}
valueBox(sit_pe[2,]$recovered, icon = "fa-user-check", color = "#ffa500")
```

### Casos Negativos
```{r}
valueBox(sit_pe[2,]$negative_tests, icon = "fa-user-minus", color = "success")
```

### Camas UCI en uso
```{r}
valueBox(beds_icu[2,]$en_uso, icon = "fa-procedures", color = "#e57373")
```

### Camas UCI disponibles
```{r}
valueBox(beds_icu[2,]$disponible, icon = "fa-bed", color = "#4dd0e1")
```


## Today

### Hoy
```{r}
valueBox(sit_pe[3,]$date, color = "#ffffff")
```

### Casos confirmados
```{r}
valueBox(sit_pe[3,]$confirmed, icon = "fa-user-plus", color = "info")
```

### Fallecimientos

```{r}
valueBox(sit_pe[3,]$deaths, icon = "fa-cross", color = "#ff0000")
```

### Casos Recuperados
```{r}
valueBox(sit_pe[3,]$recovered, icon = "fa-user-check", color = "#ffa500")
```

### Casos Negativos
```{r}
valueBox(sit_pe[3,]$negative_tests, icon = "fa-user-minus", color = "success")
```

### Camas UCI en uso
```{r}
valueBox(beds_icu[3,]$en_uso, icon = "fa-procedures", color = "#e57373")
```

### Camas UCI disponibles
```{r}
valueBox(beds_icu[3,]$disponible, icon = "fa-bed", color = "#4dd0e1")
```

## Test results and hospitalizados

```{r}
tests_raw <- read_csv("datos/covid-19-peru-test-results.csv") %>% 
  pivot_wider(
    names_from = c(resultado, tipo_prueba),
    values_from = personas
  )
last_test_date <- last(sort(unique(tests_raw$fecha)))
tests_last <- tests_raw %>% filter(fecha == last_test_date)
tests_table <- data.frame(
  res = c("Negativo", "Positivo"),
  mol = c(tests_last$negativo_moleculares, tests_last$positivo_moleculares),
  ser = c(tests_last$negativo_serológicas, tests_last$positivo_serológicas)
)
```


### Resultados de pruebas por tipo, al `r tests_last$fecha` 

```{r}
knitr::kable(
  tests_table,
  format = "html", 
  align = "rcc", 
  col.names = c("Resultado", "Pruebas moleculares", "Pruebas serológicas (\"rápidas\")"),
  format.args = list(big.mark = ",")
  ) %>% 
  kableExtra::kable_styling(
    bootstrap_options = c("striped", "bordered"),
    font_size = 32, 
    full_width = TRUE,
    
  )
```



# Nivel Nacional

## Columna unica {.tabset}

### Casos Confirmados

```{r fig.width=16, fig.height=9}
y_med <- max(covid_pe$confirmed) / 2
pe_plot <- ggplot(covid_pe,
       aes(x = date, y = confirmed)) +
  geom_point() +
  geom_line() +
  geom_vline(xintercept = as.Date("2020-03-16"), color = "blue", size = 2, alpha = .5) +
  annotate(geom = "text", x = as.Date("2020-03-16"), y = y_med, hjust = 1.1,
           label = "Inicio de la\ncuarentena", color = "blue") +
  theme_minimal(18) +
  scale_y_continuous(labels = scales::label_comma()) +
  labs(
    y = "Casos Confirmados",
    x = "",
    title = "COVID-19: Casos totales confirmados en el Perú",
    subtitle = minsa,
    caption = paste0(last_updated, " // ", jmc)
  ) +
  cur_theme

# girafe(
#   ggobj = pe_plot,
#   options = list(
#     opts_zoom(max = 5)
#   )
# )
pe_plot
```

### Trayectoria de Casos Confirmados

```{r fig.width=16, fig.height=9}
lag_days <- 14
traj_df <- covid_pe %>% 
  select(date, confirmed) %>% 
  mutate(
    new_confirmed = confirmed - lag(confirmed, lag_days)
  ) %>% 
  filter(!is.na(new_confirmed))

pelog_plot <- ggplot(traj_df,
       aes(x = confirmed, y = new_confirmed)) +
  geom_point(size = 2) +
  #geom_line() +
  geom_smooth(method = "lm", linetype = "dashed", alpha = 0.5) +
  scale_x_log10(labels = scales::comma) +
  scale_y_log10(labels = scales::comma) +
  annotation_logticks() +
  theme_minimal(18) +
  labs(
    y = paste0("Nuevos casos confirmados\n(luego de ", lag_days, " días)"),
    x = "Casos confirmados",
    title = "COVID-19: Trayectoria de los casos confirmados en el Perú",
    subtitle = minsa,
    caption = paste0(last_updated, " // ", jmc)
  ) +
  cur_theme

pelog_plot
```

### Evolución del tiempo de duplicación de casos confirmados

```{r fig.width=16, fig.height=9}
# function for running windows
dup_func <- function(df) {
  first_date = first(df$date)
  last_date = last(df$date)
  first_confirmed = first(df$confirmed)
  last_confirmed = last(df$confirmed)
  ival = (interval(first_date, last_date) %/% days())
  doubling_time = ival / log2(last_confirmed / first_confirmed)
  doubling_time
} 

# 
# pe_dt3 <- covid_pe %>% 
#   select(date, confirmed) %>% 
#   arrange(date) %>% 
#   slice(which(row_number() %% 5 == 1)) %>% 
#   mutate(
#     next_confirmed = lag(confirmed),
#     next_date = lag(date),
#     ival = (interval(date, next_date) %/% days()),
#     doubling_time =  ival / log2(next_confirmed / confirmed)
#   )

pe_dt3 <- covid_pe %>% 
  select(date, confirmed) %>% 
  arrange(date)
pe_dt3$doubling_time <- runner(
  x = pe_dt3,
  k = 14,
  f = dup_func
)

dt_eq <- "$T = \\frac{\\tau_{t} - \\tau_{t - 1}}{log_2(N_{t} / N_{t - 1})}$"

max_doubling_time = floor(max(pe_dt3$doubling_time, na.rm = TRUE))

dt_plot <- ggplot(pe_dt3, aes(x = date, y = doubling_time)) +
  #geom_line(color = "blue") +
  geom_smooth(alpha = 0.5, method = "gam") +
  geom_point() +
  geom_vline(xintercept = as.Date("2020-03-16"), color = "blue", alpha = 0.6, size = 2) +
  geom_vline(xintercept = as.Date("2020-06-30"), color = "red", alpha = 0.6, size = 2) +
  geom_vline(xintercept = as.Date("2020-07-14"), color = "grey50", linetype = "dashed", alpha = 0.6, size = 2) +
  ggrepel::geom_text_repel(aes(label = round(doubling_time, 1)), size = 3.5) +
  theme_minimal(18) +
  annotate(geom = "text", x = as.Date("2020-03-20"), size = 6, hjust = 0,
           y = 0.8*max_doubling_time, label = TeX(dt_eq)) +
  annotate(geom = "text", x = as.Date("2020-03-17"), size = 6, hjust = 0,
           color = "blue",
           y = 0.45*max_doubling_time, label = "Inicio de cuarentena") +
  annotate(geom = "text", x = as.Date("2020-07-01"), size = 6, hjust = 0,
           color = "red",
           y = 0.25*max_doubling_time, label = "Fin de cuarentena") +
  annotate(geom = "text", x = as.Date("2020-07-15"), size = 6, hjust = 0,
           color = "grey50",
           y = 0.45*max_doubling_time, label = "14 días luego del\nfin de la cuarentena") +
  labs(
    y = "Tiempo de duplicación (días)",
    x = "",
    title = "Cambio en el periodo de duplicación de casos confirmados",
    subtitle = "Evaluado usando una ventana móvil de 14 días desde el primer caso\nCurva aproximada con GAM: y ~ s(x, bs = 'cs')",
    caption = paste0(last_updated, " // ", jmc)
  ) +
  cur_theme

dt_plot
```


### Casos tamizados

```{r fig.width=16, fig.height=9}
df <- covid_pe %>% 
  rename(
    Confirmados = confirmed,
    Negativos = negative_tests
  ) %>% 
  pivot_longer(
    cols = c("Confirmados", "Negativos"),
    names_to = "Casos tamizados",
    values_to = "cases"
  ) %>% 
  select(
    date, `Casos tamizados`, cases
  )

bar_plot <- ggplot(df, 
       aes(x = date, y = cases, 
           group = `Casos tamizados`, fill = `Casos tamizados`)) +
  geom_col() +
  #geom_area() +
  theme_minimal(18) +
  labs(
    y = "Número de casos",
    x = "",
    title = "COVID-19: Casos totales tamizados en el Perú",
    subtitle = minsa,
    caption = paste0(last_updated, " // ", jmc)
  ) +
  scale_y_continuous(labels = scales::label_comma()) +
  scale_fill_manual(values = c("cyan", "lightgreen")) +
  cur_theme +
  theme(
    legend.position = c(.15,.85)
  )

bar_plot
#ggplotly(bar_plot)
```

### Casos recuperados y fallecimientos

```{r fig.width=16, fig.height=9}
recdth <- covid_pe %>% 
  select(date, recovered, deaths) %>% 
  rename(
    Recuperados = recovered,
    Fallecidos = deaths
  ) %>% 
  pivot_longer(
    cols = c(Recuperados, Fallecidos),
    names_to = "Estado",
    values_to = "cases"
  ) %>% 
  filter(!is.na(cases)) %>% 
  mutate(
    label = paste0(Estado, ": ", cases)
  )

rd_plot <- ggplot(recdth, aes(x = date, y = cases, group = Estado, fill = Estado, color = Estado)) +
  #geom_col(position = "dodge", width = 0.5) +
  geom_line(size = 2) +
  scale_y_continuous(labels = scales::comma) +
  theme_minimal(14) +
  labs(
    y = "Número de personas",
    x = "",
    title = "COVID-19: Gráficos acumulados de recuperados y fallecidos en Perú",
    subtitle = minsa,
    caption = paste0(last_updated, " // ", jmc)
  ) + 
  facet_wrap(~Estado, scales = "free_y") +
  cur_theme

# girafe(
#   ggobj = rd_plot,
#   options = list(
#     opts_zoom(max = 5)
#   )
# )
#ggplotly(rd_plot)
rd_plot
```

### Nuevos casos

```{r fig.width=16, fig.height=9}
df2 <- covid_pe %>% 
  arrange(date) %>% 
  mutate(
    new_conf = confirmed - lag(confirmed),
    new_death = deaths - lag(deaths),
    new_death = ifelse(new_death < 0, 0, new_death), # fix outliers from the start of the epidemic
    new_disc = negative_tests - lag(negative_tests),
    new_tot = new_conf + new_disc,
    new_tot = ifelse(new_tot < 0, 0, new_tot) # fix outliers from the start of the epidemic
  ) %>%
  filter(!is.na(new_conf))

# rates_df <- df2 %>%
#   mutate(
#     date_diff = as.numeric(date - lag(date)),
#     rate_conf = new_conf - lag(new_conf),
#     rate_disc = new_disc - lag(new_disc)
#   ) %>%
#   filter(!is.na(rate_conf))
# 
# ggplot(rates_df, aes(x = date, y = rate_conf)) +
#   geom_point() +
#   geom_smooth()

pt1 <- ggplot(df2, aes(x = date, y = new_conf)) +
  geom_col(fill = "darkgrey", alpha = 0.5) +
  geom_smooth(method = "gam") +
  # geom_point() +
  # geom_segment(aes(xend = date, yend = 0)) +
  theme_minimal(14) +
  labs(
    #y = "Número de casos",
    y = "",
    x = "",
    title = "Nuevos casos confirmados",
    caption = "Curva: aproximación GAM [y ~ s(x, bs = 'cs')]"
    #subtitle = minsa,
    #caption = paste0(last_updated, " // ", jmc)
  ) +
  cur_theme

pt2 <- ggplot(df2, aes(x = date, y = new_disc)) +
  geom_col(fill = "green", alpha = 0.5) +
  geom_smooth(method = "gam") +
  # geom_point(color = "blue") +
  # geom_segment(aes(xend = date, yend = 0), color = "blue") +
  theme_minimal(14) +
  labs(
    #y = "Número de casos",
    y = "",
    x = "",
    title = "Nuevos casos negativos",
    caption = "Curva: aproximación GAM [y ~ s(x, bs = 'cs')]"
    #subtitle = minsa,
    #caption = paste0(last_updated, " // ", jmc)
  ) +
  cur_theme

pt3 <- ggplot(df2, aes(x = date, y = new_death)) +
  geom_col(fill = "red", alpha = 0.5) +
  geom_smooth(method = "gam") +
  # geom_point(color = "blue") +
  # geom_segment(aes(xend = date, yend = 0), color = "blue") +
  theme_minimal(14) +
  labs(
    #y = "Número de fallecidos",
    y = "",
    x = "",
    title = "Nuevos fallecimientos",
    caption = "Curva: aproximación GAM [y ~ s(x, bs = 'cs')] // Truncado en 250"
    #subtitle = minsa,
    #aption = paste0(last_updated, " // ", jmc)
  ) +
  scale_y_continuous(limits = c(0, 250)) +
  cur_theme

pt4 <- ggplot(df2, aes(x = date, y = new_tot)) +
  geom_col(fill = "blue", alpha = 0.5) +
  geom_smooth(method = "gam") +
  # geom_point(color = "blue") +
  # geom_segment(aes(xend = date, yend = 0), color = "blue") +
  theme_minimal(14) +
  labs(
    #y = "Número de casos",
    y = "",
    x = "",
    title = "Nuevos casos tamizados",
    caption = "Curva: aproximación GAM [y ~ s(x, bs = 'cs')]"
    #subtitle = minsa,
    #caption = paste0(last_updated, " // ", jmc)
  ) +
  cur_theme

comb_plot <- (pt1 + pt2) / (pt3 + pt4)
comb_plot +
  plot_annotation(
    caption = paste0(last_updated, " // ", jmc)
  ) &
  theme(
    plot.caption = element_text(family = "inconsolata", size = 12)
  )
# 
# girafe(
#   ggobj = comb_plot,
#   options = list(
#     opts_zoom(max = 5),
#     opts_sizing(rescale = TRUE, width = .9)
#   )
# )
```

### Pacientes hospitalizados

```{r fig.width=16, fig.height=9}
hosp_raw <- read_csv("datos/covid-19-peru-detalle-hospitalizados.csv") %>% 
  arrange(fecha)
hosp_df <- hosp_raw %>% 
  select(fecha, hospitalizados, ventilacion_mecanica) %>% 
  filter(!is.na(hospitalizados) & !is.na(ventilacion_mecanica)) %>% 
  mutate(sin_ventilacion = hospitalizados - ventilacion_mecanica) %>% 
  select(-hospitalizados) %>% 
  pivot_longer(
    cols = c(ventilacion_mecanica, sin_ventilacion),
    names_to = "status",
    values_to = "count"
  ) %>% 
  mutate(
    fecha = ymd(fecha),
    status = factor(status,
                    levels = c("ventilacion_mecanica", "sin_ventilacion"),
                    ordered = TRUE)
  ) %>% 
  arrange(fecha, desc(status))

hosp_labels_df <- hosp_raw %>% 
  filter(!is.na(hospitalizados) & !is.na(ventilacion_mecanica)) %>% 
  mutate(
    fecha = ymd(fecha),
    label = paste0("Comun.\nN° ", comunicado_minsa)
  ) %>% 
  select(fecha, hospitalizados, label)

ggplot(
  data = hosp_df,
  aes(x = fecha, y = count, group = status, color = status)
) +
  geom_point() +
  geom_line(na.rm = FALSE) +
  # annotate(
  #   geom = "text",
  #   x = hosp_labels_df$fecha, 
  #   y = hosp_labels_df$hospitalizados + 100, 
  #   label = hosp_labels_df$label,
  #   size = 2
  # ) +
  scale_y_continuous(n.breaks = 10) +
  scale_x_date(date_breaks = "1 week") +
  scale_color_manual(
    values = c("red", "orange"), 
    guide = guide_legend(
      title = "Estado del paciente"
    ),
    labels = c("Con ventilación mecánica", "Sin ventilación mecánica")
  ) +
  labs(
    y = "Número de pacientes",
    x = "",
    title = "Evolución de pacientes hospitalizados por COVID-19",
    #subtitle = "Fuente: Comunicados del MINSA (indicado en el gráfico)",
    subtitle = "Fuente: Comunicados del MINSA",
    caption = paste0(last_updated, " // ", jmc)
  ) +
  theme_minimal(18) +
  cur_theme +
  theme(
    legend.position = c(.2, .8)
  ) 
```

### Casos activos

```{r fig.width=16, fig.height=9}
subset_df <- covid_pe %>%
  select(
    date,
    confirmed,
    deaths,
    recovered,
    total_tests
  ) %>%
  drop_na() %>% # get all rows with complete data
  mutate(
    active_cases = confirmed - (recovered + deaths),
    outofbounds = (active_cases < 0)
  )

ggplot(
  subset_df,
  aes(x = date, y = active_cases)
) +
  geom_point(
    aes(
      color = outofbounds,
      shape = outofbounds
    ),
    show.legend = FALSE) +
  geom_hline(yintercept = 0, linetype = "dashed") +
  ggforce::geom_mark_ellipse(
    data = subset_df %>% filter(outofbounds),
    aes(group = outofbounds),
    color = "peru"
  ) +
  scale_color_manual(
    values = c("purple", "peru")
  ) +
  scale_y_continuous(labels = scales::comma) +
  labs(
    x = "",
    y = "Número de casos",
    title = "COVID-19 - Perú: Casos activos estimados",
    subtitle = "Usando datos históricos publicados por MINSA - [Activos = Positivos - (Fallecidos + Recuperados)]\nDatos del repositorio: https://github.com/jmcastagnetto/covid-19-peru-data/",
    caption = "@jmcastagnetto, Jesus M. Castagnetto"
  ) +
  theme_minimal(18) +
  theme(
    plot.title.position = "plot",
    plot.title = element_text(size = 24, face = "bold"),
    plot.subtitle = element_text(color = "gray50"),
    plot.caption = element_text(family = "Inconsolata", size = 16)
  )

```



# Por Regiones

## mapas {.tabset data-width="40%"}

```{r echo=FALSE, out.height="90%", out.width="90%"}

days_data <- covid_regions_pe %>% 
  ungroup() %>% 
  select(date) %>% 
  distinct() %>% 
  arrange(desc(date)) %>% 
  pull(date)
#days_selected <- days_data[seq(1, length(days_data), 4)]
days_selected <- days_data[1:15] #últimos 15 días

last_date <- days_data[1]

map_df <- covid_regions_pe %>% 
  ungroup() %>%
  filter(date %in% days_selected) %>% 
  select(region, date, confirmed, confirmed_per_million) %>% 
  mutate(
    region = str_to_title(region) %>% 
      str_replace("Huanuco", "Huánuco") %>% 
      str_replace("Junin", "Junín") %>% 
      str_replace("De Dios", "de Dios") %>% 
      str_replace("Martin", "Martín") %>% 
      str_replace("Apurimac", "Apurímac")
  )

last_date_index <- map_df %>% 
  select(date) %>% 
  distinct() %>% 
  nrow()

map1_max <- ((max(covid_regions_pe$confirmed) %/% 1000) + 1) * 1000

rmap_1 <- map_df %>% 
  group_by(date) %>% 
  e_charts(region, timeline=TRUE) %>%
  e_timeline_opts(currentIndex = last_date_index - 1) %>% 
  em_map("Peru") %>% 
  e_map(confirmed, map = "Peru", roam = TRUE) %>% 
  e_visual_map(min = 1, max = map1_max) %>% 
  e_tooltip(formatter = e_tooltip_choro_formatter("decimal")) %>% 
  e_title(
    text = "Número de casos confirmados por región",
    subtext = paste0("Últimos 15 días (hasta el ", last_date, ")")
  ) %>% 
  e_theme("infographic") %>% 
  e_toolbox() %>% 
  e_toolbox_feature(feature = "saveAsImage", title = "Guardar como imágen") #%>% 
  #e_toolbox_feature(feature = "restore", title = "Restaurar mapa") %>% 
  #e_group("regiones")

# map_df2 <- covid_regions_pe %>% 
#   select(region, date, confirmed_per_million) %>% 
#   mutate(
#     region = str_to_title(region) %>% 
#       str_replace("Huanuco", "Huánuco") %>% 
#       str_replace("Junin", "Junín") %>% 
#       str_replace("De Dios", "de Dios") %>% 
#       str_replace("Martin", "Martín") %>% 
#       str_replace("Apurimac", "Apurímac")
#   )

map2_max <- ((max(map_df$confirmed_per_million) %/% 100) + 1) * 100

rmap_2 <- map_df %>% 
  ungroup() %>% 
  group_by(date) %>% 
  e_charts(region, timeline=TRUE) %>%
  e_timeline_opts(currentIndex = last_date_index - 1) %>% 
  em_map("Peru") %>% 
  e_map(confirmed_per_million, map = "Peru", roam = TRUE) %>% 
  e_visual_map(min = 1, max = map2_max) %>% 
  e_tooltip(formatter = e_tooltip_choro_formatter("decimal")) %>% 
  e_title(
    text = "Densidad de casos confirmados por millón en cada región",
    subtext = paste0("Últimos 15 días (hasta el ", last_date, ")")
  ) %>% 
  e_theme("infographic") %>% 
  e_toolbox() %>% 
  e_toolbox_feature(feature = "saveAsImage", title = "Guardar como imágen") #%>% 
  #e_toolbox_feature(feature = "restore", title = "Restaurar mapa") %>% 
  #e_group("regiones") #%>% 
  #e_connect_group("regiones")

#e_arrange(rmap_1, rmap_2, rows = 1, cols = 2)
```

### Densidad (casos por millón de habitantes) por región

```{r}
rmap_2
```

### Número de casos por región

```{r}
rmap_1
```

## plots {.tabset data-width="60%"}

### Evolución de casos por región

```{r fig.height=9, fig.width=14}
cpr_df <- covid_regions_pe %>% 
  mutate(
    label = glue::glue("{region}: {confirmed} ({date})")
  )
cpr <- ggplot(cpr_df,
  aes(x = date, y = confirmed, color = region)) +
  geom_point(show.legend = FALSE, size = 0.5) +
  #geom_segment(aes(xend = date, yend = 0), size = 1, show.legend = FALSE) +
  geom_line(show.legend = FALSE, size = 0.5) +
  ylim(0, NA) +
  facet_wrap(~region, scales = "free_y", ncol = 5) +
  theme_minimal(10) +
  labs(
    y = "Casos Confirmados",
    x = "",
    title = "COVID-19: Casos Confirmados por region en el Perú",
    subtitle = minsa,
    caption = paste0(last_updated, " //  ", jmc)
  ) +
  cur_theme
cpr
#ggplotly(cpr)
```

### Casos positivos por tipo de prueba

```{r fig.height=9, fig.width=14}
tests_reg <- covid_regions_pe %>% 
  select(region, date, 
         pcr_test_positive, 
         serological_test_positive,
         pcr_serological_test_positive) %>% 
  pivot_longer(
    cols = c(pcr_test_positive, 
         serological_test_positive,
         pcr_serological_test_positive),
    names_to = "tipo_prueba",
    values_to = "cantidad", 
    values_drop_na = TRUE
  ) %>% 
  mutate(
    tipo_prueba = factor(
      tipo_prueba, 
      levels = c("pcr_test_positive", 
         "serological_test_positive",
         "pcr_serological_test_positive"),
      ordered = TRUE
    )
  )

ggplot(
  tests_reg,
  aes(x = date, y = cantidad, group = tipo_prueba, fill = tipo_prueba)
) +
  geom_col() +
  scale_fill_manual(
    values = c("cyan", "green", "peru"), 
    guide = guide_legend(
      title = "Tipo de prueba"
    ),
    labels = c("PCR (molecular)",
               "Serológica (\"rápida\")",
               "PCR + Serológica")
  ) +
  labs(
    y = "",
    x = "",
    title = "Pruebas positivas por tipo en cada región",
    caption = paste0(last_updated, " //  ", jmc)
  ) +
  facet_wrap(~region, scales = "free_y", nrow = 5) +
  theme_classic(14) +
  cur_theme +
  theme(
    legend.position = "top"
  )
  
```


# Datos

## Column {.tabset}

### Nivel Nacional

```{r results="asis"}
tab_opts <- list(
    language = list(
      search = "Filtrar:",
      url = '//cdn.datatables.net/plug-ins/1.10.11/i18n/Spanish.json'
      ),
    className = "dt-center",
    searchHighlight = TRUE,
    pageLength = 15,
    lengthMenu = c(15, 30, 40, 50),
    dom = 'Blfrtip',
    buttons = c('csv', 'excel')
  )

# ref: https://www.worldometers.info/world-population/peru-population/
# retrieved: 2020-04-09
peru_pop <- 32865689 

dt_df <- covid_pe %>% 
  mutate(
    confirmed_per_million = round(confirmed * 1e6 / peru_pop, 1)
  ) %>% 
  select(
    country,
    iso3c,
    date,
    confirmed,
    confirmed_per_million,
    pcr_test_positive,
    serological_test_positive,
    ag_test_positive,
    pcr_serological_test_positive,
    deaths,
    recovered,
    negative_tests,
    total_tests
  ) %>% 
  arrange(desc(date))

datatable(
  dt_df,
  colnames = c("País", 
               "Código ISO 3166-1 alpha 3", 
               "Fecha", 
               "Número de casos confirmados", 
               "Casos confirmados por millón de personas",
               "Positivos por pruebas moleculares",
               "Positivos por pruebas serológicas",
               "Positivos por pruebas antigénicas",
               "Positivos por moleculares y serológicas",
               "Número de fallecimientos",
               "Número de casos recuperados",
               "Número de casos negativos",
               "Número de pruebas totales"
               ),
  class = "cell-border stripe",
  extensions = c("Buttons", "Responsive"),
  options = tab_opts,
  caption = paste0(minsa, " // ", last_updated, " // ", jmc)
)
```


### Por región

```{r results="asis"}
datatable(
  covid_regions_pe %>% 
    select(
      region,
      ubigeo,
      iso_3166_2_code,
      date,
      confirmed,
      confirmed_per_million,
      pcr_test_positive,
      serological_test_positive,
      ag_test_positive,
      pcr_serological_test_positive,
      deaths,
      negative_tests,
      total_tests,
      pop2020
    ) %>% 
    arrange(desc(date), region),
  colnames = c("Departamento", 
               "UBIGEO (INEI)",
               "Código ISO 3166-2:PE",
               "Fecha", 
               "Número de casos confirmados",
               "Casos confirmados por cada millón de personas",
               "Positivos por pruebas moleculares",
               "Positivos por pruebas serológicas",
               "Positivos por pruebas antigénicas",
               "Positivos por moleculares y serológicas",
               "Número de fallecimientos",
               "Pruebas negativas",
               "Total de pruebas",
               "Población (2019)"
               ),
  class = "cell-border stripe",
  extensions = c("Buttons", "Responsive"),
  options = tab_opts,
  caption = paste0("** ", no_deaths_minsa_20200412, " ** ", 
                   minsa, " // ", last_updated, " // ", jmc)
)
```

# Acerca de esta visualización

## Columna única

- **Fuente**: Tweets del MINSA https://twitter.com/Minsa_Peru
- **Datos y código**: https://github.com/jmcastagnetto/covid-19-peru-data/

Esta visualización esta hecha usando el lenguaje de programación [R](https://www.r-project.org/),
y hace uso de una serie de librerías ([flexdashboard](https://rmarkdown.rstudio.com/flexdashboard), [ggplot2](https://ggplot2.tidyverse.org/), [echarts4r](https://echarts4r.john-coene.com/), [DT](https://rstudio.github.io/DT/)), para producir los distintos gráficos y tablas.

[`r fontawesome::fa("twitter")`](https://twitter.com/jmcastagnetto)
[`r fontawesome::fa("github")`](https://github.com/jmcastagnetto')
[`r fontawesome::fa("home")`](https://castagnetto.site')
Jesús M. Castagnetto, Ph.D.