---
title: "Changing funnel plot options"
output: 
  rmarkdown::html_vignette:
    toc: true
    toc_depth: 2
vignette: >
  %\VignetteIndexEntry{Changing funnel plot options}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
  \usepackage[utf8]{inputenc}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.height = 5, 
  fig.width = 7
)
```

## Setup

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

This brief vignette shows how to go about changing some of the parameters available in the `FunnelPlotR` package.
Firstly, lets set up some data, the same as the README and other vignette:

```{r dtsetup}
library(COUNT)
data(medpar)
medpar$provnum<-factor(medpar$provnum)
medpar$los<-as.numeric(medpar$los)

# Logistic model to predict LOS, LOS is quite overdispersed
mod<- glm(los ~ hmo + died + age80 + factor(type), family="poisson", data=medpar)

#Get predicted value for ratio
medpar$prds<- predict(mod, newdata = medpar, type="response")

# Draw plot, returning just the plot object
funnel_plot(medpar, denominator=prds, numerator=los
            , group = provnum, limit=99 ,label = "outlier"
            , draw_unadjusted = TRUE)

```


## Highlighting a data point

You can pick out data point(s) using the `highlight` option.  Here we we use the example above to highlight the hospital labels '030002' in the data set.

```{r highlight}
# Draw plot, returning just the plot object
funnel_plot(medpar, denominator=prds, numerator=los
            , group = provnum, limit=99 ,label = "outlier"
            , draw_unadjusted = TRUE, highlight="030002")



```


## Themes

You can alter themes in the `FunnelPlotR` packages by using the theme argument.  There are a couple of options included with the package `funnel_clean` (the default) and `funnel_grey`, but you can write your own theme using any valid `ggplot2` theme and pass it to the plot.

```{r plottheme1}

funnel_plot(medpar, denominator=prds,numerator=los
            , group = provnum, limit=99 ,label = "outlier"
            , draw_unadjusted = TRUE, theme = funnel_grey() )

```

If you are not familiar with themes, you can create a theme with the `theme` function from `ggplot2`. I will create a new theme, including the `funnel_grey` theme, but with some different (exaggerated) elements.

```{r plottheme2}
library(ggplot2)

new_funnel_theme <-
  funnel_grey()+
    theme(plot.title = element_text(face="bold", colour="red", size=6), # Change plot title
          legend.background = element_rect(fill="brown"), # Alter legend background colour
          axis.title.y = element_text(angle=0)  #Rotate y axis label
    )


funnel_plot(medpar, denominator=prds,numerator=los
            , group = provnum, limit=99 ,label = "outlier"
            , draw_unadjusted = TRUE, theme = new_funnel_theme)

```

## Plot colours

You can change the colours of the limits.  I am aiming to allow better colour theme support in future versions, but at present they are supplied as a vector of eight hex colours to the `plot_cols` argument, in the order 95% Poisson (lower/upper), 99.8% Poisson (lower/upper), 95% OD (lower/upper) and 99.8% OD (lower/upper).  At present all eight values are required, whether you are plotting all eight limits or not.
I have deliberately avoided using red and green colours as defaults because it encourages value judgements about 'good' v.s. 'bad' which may be unreasonable for a given data set.  Default colours are paired for both low and high: `c("#FF7F0EFF", "#FF7F0EFF", "#1F77B4FF","#1F77B4FF", "#9467BDFF", "#9467BDFF", "#2CA02CFF", "#2CA02CFF")`

Here I will change the upper 95% Poisson limit to black ("#000000"):

```{r colours}
funnel_plot(medpar, denominator=prds,numerator=los
            , group = provnum, limit=99 ,label = "outlier"
            , draw_unadjusted = TRUE, theme = funnel_grey(),
            plot_cols = c("#FF7F0EFF", "#000000", "#1F77B4FF","#1F77B4FF", "#9467BDFF", "#9467BDFF", "#2CA02CFF", "#2CA02CFF"))
```

## Changing scales

`FunnelPlotR` automatically sets it's scales from the data you present to it but, on occasions, the scale rules might fall down for a particular dataset, or you may want to fix the plot to a particular scale.  You can do this using the `xrange` and `yrange` arguments.  Each takes a vector of two values, the minimum and the maximum:

```{r funnelscales}
## Changing labels
funnel_plot(medpar, denominator=prds,numerator=los
            , group = provnum, limit=99 ,label = "outlier"
            , draw_unadjusted = TRUE, x_range=c(0, 400), y_range=c(0,2))

```

`ggplot2` will warn you, as above, if you are excluding any points from your plot.

## Changing labels

You can change the plot labels and axis labels easily using the options: `title`, `x_label` and `y_label`.

```{r funnellabels1}
funnel_plot(medpar, denominator=prds,numerator=los
            , group = provnum, limit=99 ,label = "outlier"
            , draw_unadjusted = TRUE, title = "Vignette funnel plot"
            , x_label = "x-axis", y_label = "y-axis")
```

There are different labelling options for the data points too, using the `label` option.  The default is to label outliers, but you can turn labels off, label the highlighted points, or both the highlighted points and the outliers ('both')

```{r funnellabels2}
funnel_plot(medpar, denominator=prds,numerator=los
            , group = provnum, limit=99
            , draw_unadjusted = TRUE, title = "Vignette funnel plot"
            , x_label = "x-axis", y_label = "y-axis"
            , highlight= "030002", label = "highlight")
```

## Cutting out the ggplot object

Since `FunnelPlotR` uses `ggplot2`, you could always extract the plot and alter it manually like any other `ggplot2` object.  The easiest way is to extract it with `plot()`.  Below we'll add a (completely useless) vertical line to demonstrate adding more elements:

```{r cutoutplot}
# Original funnel plot object
fp <-
funnel_plot(medpar, denominator=prds,numerator=los
            , group = provnum, limit=99, label = "outlier"
            , draw_unadjusted = TRUE)

# Extract just the plot
my_plot <- plot(fp)

# Add an additional geom to plot
my_plot +
  geom_vline(aes(xintercept=400), linetype = "dashed", colour="red", linewidth=2)

```