Check Turkey underascertainment estimate

[adamkucharski]

In the estimating underascertainment vignette, the estimate for Turkey is very narrow (see below). Need to check whether this is just a feature of the data, or whether indicative of an edge case.

[adamkucharski]

The Turkey estimate is valid, just with very small CI. This is down to the decline of the epidemic, and hence relatively few cases with outcomes not yet known. Below is a comparison of several countries:

Code:

# Plot data
covid_turkey <- covid_data_early |> filter(country=="Turkey")

p1 <- ggplot(covid_turkey) +
  geom_line(
    aes(x = date, y = cases),
    colour = "darkblue"
  ) +
  geom_line(
    aes(x = date, y = deaths), 
    colour = "darkred"
  ) +
  ggtitle("Turkey")

covid_india <- covid_data_early |> filter(country=="India")

p2 <- ggplot(covid_india) +
  geom_line(
    aes(x = date, y = cases),
    colour = "darkblue"
  ) +
  geom_line(
    aes(x = date, y = deaths), 
    colour = "darkred"
  ) +
  ggtitle("India")

covid_russia <- covid_data_early |> filter(country=="Russia")

p3 <- ggplot(covid_russia) +
  geom_line(
    aes(x = date, y = cases),
    colour = "darkblue"
  ) +
  geom_line(
    aes(x = date, y = deaths), 
    colour = "darkred"
  ) +
  ggtitle("Russia")

covid_poland<- covid_data_early |> filter(country=="Poland")

p4 <- ggplot(covid_poland) +
  geom_line(
    aes(x = date, y = cases),
    colour = "darkblue"
  ) +
  geom_line(
    aes(x = date, y = deaths), 
    colour = "darkred"
  ) +
  ggtitle("Poland")

cowplot::plot_grid(p1,p2,p3,p4, nrow = 2, ncol = 2)