local_profile.Rmd

---
title: ""
output: pdf_document
geometry: margin=0.5cm
params:
  reactive_df: NA
  chosen_area: NA
  chosen_sex: NA
  chosen_geography_level: NA
header-includes:
   - \usepackage[default]{sourcesanspro}
   - \usepackage[T1]{fontenc}
   - \usepackage{fancyhdr}
   - \usepackage{graphicx}
   - \usepackage{longtable}
   - \usepackage{tabu}
   - \usepackage{booktabs}
   - \usepackage{array}
   - \usepackage{multirow}
   - \usepackage{wrapfig}
   - \usepackage{float}
   - \usepackage{colortbl}
   - \usepackage{pdflscape}
   - \usepackage{threeparttable}
   - \usepackage{threeparttablex}
   - \usepackage[normalem]{ulem}
   - \usepackage{makecell}
   - \usepackage{xcolor}   
   - \usepackage{hyperref}
   - \newcommand{\blandscape}{\begin{landscape}}
   - \newcommand{\elandscape}{\end{landscape}}
mainfont: SourceSansPro


---
```{r, include=FALSE}
options(tinytex.verbose = TRUE)
```

```{r echo=FALSE, message= FALSE}

library(ggplot2)
library(dplyr)
library(kableExtra)
library(tidyr)
library(knitr)
library(stringr)

chosen_area <- gsub("&", "\\\\&", params$chosen_area)

mydate <- Sys.Date()

```


\blandscape

  \begin{minipage}{0.7\textwidth}
    \raggedright
    \textbf{\large{PHS Adult Mental Health Indicators}} \\
    \large{Local area profile for `r chosen_area` (`r params$chosen_geography_level`)}\\
    Sex: `r params$chosen_sex` \\
    Downloaded on: `r mydate` \\
    From: Public Health Scotland's Adult Mental Health Indicators dashboard \\
  \end{minipage}
  \hfill
  \begin{minipage}{0.3\textwidth}
    \raggedleft
    \includegraphics[width=100pt]{phs-logo-updated.png}
  \end{minipage}
  \bigskip
  \begin{minipage}{0.8\textwidth}
    \raggedright
    \bigskip
    \large{This local profile compares the latest available indicator data for the selected area (`r chosen_area`) with data for the other `r params$chosen_geography_level`s and the Scottish average.} \par 
    \bigskip
    \large{The grey 'spine chart' bar shows the range of values for each indicator, and the circle shows where the selected area places within the range, as shown in this image. The red line indicates the Scottish average. The colour of the circle indicates whether the value in the selected area is significantly 'worse' than the Scottish average (yellow), significantly 'better' (blue), or not significantly different (grey). Data for individual indicators can be viewed and downloaded on the dashboard.} \par 
    \bigskip
    \large{The profile is intended to increase understanding of local health issues and to prompt further investigation, rather than to be used as a performance assessment or benchmarking tool. Local knowledge is needed to understand and interpret differences, as there can be many reasons why an indicator is higher in one area compared to another.} 
  \end{minipage}
  \begin{minipage}{0.3\textwidth}
    \raggedleft
    \includegraphics[width=250pt]{spine.png}
    \includegraphics[width=250pt]{local_profile_legend.png}
  \end{minipage}



```{r setup, include=FALSE, dev="cairo_pdf"} 
knitr::opts_chunk$set(warning = FALSE, message = FALSE) 
knitr::opts_chunk$set(fig.pos = "H", out.extra = "")
```


```{r echo=FALSE, message= FALSE}

# function to create spine plot for each indicator
create_spine_plot <- function(data, chosen_area_colour) {
  plot <- ggplot(data, aes(x = ind_name, y = GraphPoint, fill = fct_rev(intervals))) + # intervals need reversing to plot correctly
    geom_bar(stat = "identity", width = 0.75) +
    scale_fill_manual(values = c("int1" = 'white', "int2" = '#D3D3D3', "int3" = '#A4A4A4', "int4" = '#D3D3D3', "int5" = 'white')) +
    geom_hline(yintercept = 0.5, col = "red") +
    ylim(0, 1) +
    geom_point(data = NULL, aes(x = ind_name, y = new_chosen_value), shape = 19, size = 4, col = chosen_area_colour) +
    coord_flip() +
    theme_void() +
    guides(fill = "none") +
    theme(plot.margin = unit(c(0, 0, 0, 0), "cm"),
          plot.background = element_rect(fill = "white", color = "white"),
          panel.background = element_rect(fill = "white", color = "white")
          )

  return(plot)
}

# function to create sparkline for each indicator
create_sparkline <- function(data) {
  plot <- ggplot(data, aes(x=1:length(trend), y=as.numeric(trend))) +
    geom_line(linewidth=2, colour='#0078D4') +
    geom_area(fill="#E6F2FB") +
    theme_void() +
    guides(fill = "none") +
    scale_y_continuous(limits=c(0, NA), expand = expansion(mult = c(0, 0.2))) +
    theme(plot.margin = unit(c(0, 0, 0, 0), "cm"),
          plot.background = element_rect(fill = "white", color = "white"),
          panel.background = element_rect(fill = "white", color = "white"))
  
  return(plot)
}




# temporary folder to save charts to
graph_directory <- tempdir()



# # pivot reactive dataframe in longer format (required for ggplot)
df <- params$reactive_df %>%
   select(domain, ind_name, short_definition, year_range, int1, int2, int3, int4, int5, scotland = scotland_value, value, new_chosen_value, marker_colour, trend) %>%
   pivot_longer(cols = c(int1, int2, int3, int4, int5), values_to = "GraphPoint", names_to = "intervals") %>%
   # order quantiles so stack in correct order
   mutate(intervals = factor(intervals, levels = c("int1", "int2", "int3", "int4", "int5")))



# the data to be displayed in table
table <- df %>%
  select(domain, ind_name, short_definition, value, scotland, year_range, trend) %>%
  distinct() %>%
  mutate(ind = paste0(
    "\\textbf{",ind_name, "}\n", "\\textit{", short_definition, " ", year_range, ".}")) %>%
  mutate(id = row_number(),
         spine ="") %>% #placeholder, so there are enough cols
  select(domain, ind, id, value, scotland, spine) 

  


table$ind <- linebreak(table$ind)



# get unique indicator list
 unique_indicators <- unique(df$ind_name)

 # Counter variable to keep track of which chart being generated
counter <- 1

# Iterate over each indicator and create/save a spineplot

for (ind_name in unique_indicators) {

  # Filter the data for the current indicator
  indicator_data <- df %>% filter(ind_name == !!ind_name)

  # get colour for point
  chosen_area_colour <- unique(indicator_data$marker_colour)

  # create the plot
  plot <- create_spine_plot(indicator_data, chosen_area_colour)

  # create filename
  filename <- sprintf("%s/%s.png", graph_directory, counter)

  # Save the plot as a PNG with the filename created
  ggsave(filename, plot = plot, units="mm", width=50, height=5, dpi=300)
  
  # Unnest the trend column for plotting
  indicator_data <- indicator_data[1,] %>% # 1st row, as contents of trend column will be repeated for the same indicator
    select(trend) %>%
    unnest(cols = c(trend))
 
  # create the plot
  plot2 <- create_sparkline(indicator_data)
  
  # create filename
  filename2 <- sprintf("%s/spark%s.png", graph_directory, counter)
  
  # Save the plot as a PNG with the filename created
  ggsave(filename2, plot = plot2, units="mm", width=15, height=5, dpi=300)

  # increase counter variable
  counter <- counter + 1
}

# define column headers for the pdf table
col_names = c("Domain", "Indicator", "Time trend", 
              chosen_area, "Scotland", 
              "Spine chart"
              )

```



```{r, echo= FALSE, message = FALSE}

 # the final table to be displayed with charts embedded

table %>%
  group_by(domain) %>%
  knitr::kable(
    format = "latex", # must be in latex format for pdf download
    booktabs = TRUE, # generic formatting
    escape = FALSE, # ensures latex formatting on indicator column works (i.e.bold/italics,linebreaks)
    longtable = TRUE, # ensures table goes across multiple pages when too long
    align = c("llllll"), #left alignment within cells
    col.names = col_names
     ) %>% 
  
  # format header
  row_spec(0, background = "#032F4F", color = "white") %>%
  add_header_above(., c(" " = 3,
                        "Latest values" = 3),
                   escape = TRUE,
                   background = "#032F4F", color = "white") %>%
  # set column widths
   column_spec(1, width = "8em") %>% # domain column
   column_spec(2, width = "20em") %>% # indicator column
    # embed chart images on each row of column 6
   kableExtra::column_spec(3, width = "6em",
                           image = sprintf("%s/spark%s.png", graph_directory, table$id), color = "white") %>%
   column_spec(4, width = "6em") %>% # chosen area column
   column_spec(5, width = "3em") %>% # scotland column

  # embed chart images on each row of column 5
   kableExtra::column_spec(6, width = "20em",
                           image = sprintf("%s/%s.png", graph_directory, table$id), color = "white") %>%
  

  # repeats column headings when table goes across multiple pages
   kable_styling(latex_options = c("repeat_header")) %>%

  # add footnote with link to profiles tool
  footnote(general="\\\\url{https://scotland.shinyapps.io/phs-adult-mhi-dashboard/}",
           general_title = "Source: ", footnote_as_chunk = T, escape=F) %>%

  # prevent domain from appearing on every row of column 
  kableExtra::collapse_rows(columns = 1, valign = "top", row_group_label_position = "first", custom_latex_hline = 2:6, longtable_clean_cut = F)


```

\elandscape