---
title: "Estimate asymptomatic cases in Italy during the COVID-19 pandemic"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Estimate asymptomatic cases in Italy during the COVID-19 pandemic}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

```{r setup}
library(asymptor)
```

Let's start by loading the example data. It's bundled in the package but 
originally comes from https://github.com/GoogleCloudPlatform/covid-19-open-data
(Apache License 2.0).

```{r}
df <- readRDS(system.file("extdata", "covid19_italy.rds", package = "asymptor"))
head(df)
```

We can feed this data directly to the `estimate_asympto()` function. This 
function requires 3 columns, labelled as `date`, `new_cases`, `new_deaths`,
containing the daily counts (not the cumulated total!)

```{r}
asy <- estimate_asympto(df$date, df$new_cases, df$new_deaths)
head(asy)
```

We may want to visualise these estimations alongside the empirical data. So,
we start by merging the two datasets:

```{r}
res <- merge(df, asy)
head(res)
```

Alternatively, we can directly use a tidyverse-compatible syntax:

```{r, eval = require("dplyr")}
library(dplyr)
res <- df %>%
  mutate(lower = estimate_asympto(date, new_cases, new_deaths, "lower")$lower,
         upper = estimate_asympto(date, new_cases, new_deaths, "upper")$upper)
head(res)
```
Then, we can the ggplot2 package to plot the result:

```{r, example_fig, fig.height = 4.5, fig.width = 9, out.width='100%'}
library(ggplot2)
ggplot(res, aes(x = date)) +
  geom_line(aes(y = new_cases+lower), col = "grey30") +
  geom_ribbon(aes(ymin = new_cases+lower, 
                  ymax = new_cases+upper), 
              fill = "grey30") +
  geom_line(aes(y = new_cases), color = "red") +
  labs(title = "Estimated total vs detected cases of COVID-19 in Italy",
       y = "Cases") +
  theme_minimal()
```