README.Rmd

---
output: github_document
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)
```

# edr

## Estimated Dissemination Ratio


This package is a fast implementation of the Estimated Dissemination Ratio, 
described here

[Estimated Dissemination Ratio—A Practical Alternative to the Reproduction Number for Infectious Diseases](https://dx.doi.org/10.3389/fpubh.2021.675065)

## Installation

``` {r,eval=FALSE}
# install.packages("devtools")
devtools::install_github("Nic-Chr/edr")
```

### Libraries

``` {r}
library(edr)
library(data.table)
library(ggplot2)
```

### Recreating Figure 1

To recreate figure 1 from the paper, we use data from the John Hopkins COVID-19 data repository.    
Source: [CSSEGISandData/COVID-19](https://github.com/CSSEGISandData/COVID-19)

### Data

``` {r}
data(uk_covid_cases)
setDT(uk_covid_cases)
```

### Some technical notes

`edr()` returns a `data.table` with class 'edr'

```{r,error=TRUE}
temp <- edr(uk_covid_cases$new, 7)
temp
class(temp)
```

If you want just the edr estimates as a regular vector you can use `edr_only()`

```{r}
temp[, edr2 := edr_only(cases, 7)][]

rm(temp) # Remove temp
```


### EDR calculation 

Here we calculate the 7-day EDR with `edr()` 
along with 99% percentile confidence intervals

``` {r}
edr_seven_day <- uk_covid_cases[, edr(new, window = 7, order_by = reporting_date,
                                     simulations = 1e04, alpha = 0.01)]
```

We also calculate the 7-day rolling average of new confirmed cases

``` {r}
edr_seven_day[, ma7 := frollmean(cases, n = 7, align = "right")][]
```

Finally plotting everything

``` {r}
edr_seven_day <- edr_seven_day[
  time >= as.Date("2020-03-10") & 
    time < as.Date("2020-05-01")]

scale_factor <- 500
uk_lockdown <- as.Date("2020-03-24")
```

There is a convenient plot method for edr objects defined in `plot.edr()`.

``` {r}
edr_plot <- plot(edr_seven_day, include_cases = FALSE)
edr_plot
```

We just need to add a few things to recreate the figure

```{r}
edr_plot +
  geom_col(aes(y = ma7 / scale_factor), width = 1, alpha = 0.4, fill = "#0077B6") +
  geom_segment(aes(x = uk_lockdown, y = 6, xend = uk_lockdown, yend = 3.25),
               arrow = arrow(length = unit(0.5, "cm"))) +
  annotate("text", x = uk_lockdown, y = 6, 
           label = "full UK lockdown \nimplemented",
           vjust = -0.1,
           fontface = "bold") +
  labs(x = "Reporting Date", y = "UK EDR") +
  scale_y_continuous(breaks = seq(0, 10, 1),
                     sec.axis = sec_axis(\(x) x * scale_factor,
                                         breaks = seq(0, 5000, 500),
                                         name = "UK new case 7-day rolling average")) +
  theme_bw() + 
  theme(axis.line.y.right = element_line(color = "#0077B6"), 
        axis.ticks.y.right = element_line(color = "#0077B6"),
        axis.text.y.right = element_text(color = "#0077B6"), 
        axis.title.y.right = element_text(color = "#0077B6")
  )
```