Load the functions and data we'll use

worked_example/app.R

@@ -8,6 +8,11 @@
 #
 
 library(shiny)
+library(ggplot2)
+library(dplyr)
+
+gapminder <- readRDS("gapminder.rds")
+source("workshopFunctions.R")
 
 # Define UI for application that draws a histogram
 ui <- fluidPage(

worked_example/workshopFunctions.R

@@ -0,0 +1,168 @@
+# Functions used in the Workshop
+
+#' Plot a single year of gapminder data, passed in indata.
+#' makeSemiTransparent will make the points semi transparent;
+#' this  produces a clearer graph if we overlay historic data on it
+produceGapminderPlot <- function(indata, 
+                                 makeSemiTransparent = FALSE,
+                                 x = gdpPerCap,
+                                 y = lifeExp){
+  x <- ensym(x)
+  y <- ensym(y)
+
+  # These are taken from the gapminder package, but made brighter
+  continent_colours <-c(Africa = "#BF590CFF", Americas = "#F80039FF", Asia = "#600071FF", 
+                        Europe = "#3B9626FF", Oceania = "#4A51E0FF")
+
+  # Set nice axes for the data
+  # We can't do this directly from the data since the limits would change
+  # for each year.
+  axislimits <- list(tfr = c(0.5, 9),
+                     lifeExp = c(20, 88),
+                     gdpPerCap = c(250, 60000))
+
+  axislabels <- list(tfr = "Total Fertility Rate",
+                     lifeExp = "Life Expectancy (years)",
+                     gdpPerCap = "GDP per capita (ppp dollars)")
+
+  if(!all(names(axislimits) %in% names(axislabels)) || 
+     !all(names(axislabels) %in% names(axislimits)) ){
+    stop("Haven't defined the axis limits and labels for the same variables")
+  }
+
+  if(!(toString(x) %in% names(indata))) {
+    stop(paste(toString(x), "not found in input data"))
+  }
+
+  if(!(toString(y) %in% names(indata))) {
+    stop(paste(toString(y), "not found in input data"))
+  }
+
+  if(!(toString(x) %in% names(axislimits))) {
+    stop(paste("Don't know how to handle the axis/labels for", toString(x)))
+  }
+
+  if(!(toString(y) %in% names(axislimits))) {
+    stop(paste("Don't know how to handle the axis/labels for", toString(y)))
+  }
+
+  plotYear <- min(indata$year)
+
+  if (nrow(indata) > 0 && max(indata$year) != plotYear) {
+    stop("More than one year of data passed to plotting function - use filter(year == xxxx)")
+  }
+
+
+  # Set up the plot window, with appropriate limits
+  gapminderPlot <- ggplot(data = indata) +
+    coord_cartesian(xlim = axislimits[[toString(x)]],
+                    ylim = axislimits[[toString(y)]]) 
+
+  # We need to log scale gdpPerCap
+  if(toString(x) == "gdpPerCap"){
+    gapminderPlot <- gapminderPlot +
+      scale_x_log10(breaks = 4*10**(2:4)) 
+  }
+
+  if(toString(y) == "gdpPerCap"){
+    gapminderPlot <- gapminderPlot +
+      scale_y_log10(breaks = 4*10**(2:4)) 
+  }
+
+  # Return the empty plot if we've no data
+  if (nrow(indata) == 0){
+    return(gapminderPlot)
+  }
+
+
+  if (makeSemiTransparent) {
+    pointAlpha <- 0.3
+  } else {
+    pointAlpha <- 1
+  }
+
+  ## Plot the data
+  gapminderPlot <- gapminderPlot +  
+    geom_point(aes(x = !!x,
+                   y = !!y, 
+                   colour = continent,
+                   size = population),
+               alpha = pointAlpha) + 
+    labs(title = plotYear,
+         x = axislabels[[toString(x)]],
+         y = axislabels[[toString(y)]],
+         colour = "Continent") +
+    scale_colour_manual(values = continent_colours) +
+    scale_alpha_manual() +
+    guides(size = FALSE) 
+
+  return(gapminderPlot)
+
+}
+
+
+# Get the variable name plotted on an axis from a ggplot object
+# We use this function to make sure we plot the historic data for the
+# same variables as the main plot
+getAxis <- function(g, axis) {
+
+  if( length(g$layers) > 1 ){
+    stop("Only implemented for single layer graphs")
+  } 
+
+  if ( !(axis %in% c("x","y")) ) {
+    stop("Invalid axis specified")
+  }
+
+  return(rlang:::get_expr(g$layers[[1]]$mapping[[axis]]))
+
+}
+
+# Add historic data to a "gapminder" plot.
+# This function expects a single country's data
+addHistoricPlot <- function(inplot, indata, activeYear = NULL){
+
+  if (length(unique(indata$country)) > 1 ) {
+    stop("More than one county's data passed")
+  }
+
+  if (is.null(activeYear)) {
+    activeYear = max(indata$year)  
+  }  
+
+  xdata = getAxis(inplot, "x")
+  ydata = getAxis(inplot, "y")
+
+  outplot <- inplot + geom_path(data = indata,
+                                aes(x = !!xdata,
+                                    y = !!ydata,
+                                    colour = continent),
+                                show.legend = FALSE) +
+    geom_point(data = indata[indata$year == activeYear,],
+               aes(x = !!xdata,
+                   y = !!ydata,
+                   colour = continent,
+                   size = population),
+               show.legend = FALSE)
+
+  return(outplot)
+
+}
+
+# get the richest country in a tibble, and return
+# the country and its GDP as a tibble
+getRichestCountry <- function(indata) {
+
+  if (max(indata$year) != min(indata$year) ) {
+    warning("More than one year of data passed in")
+  }
+
+  outdata <- indata %>% 
+    arrange(desc(gdpPerCap)) %>% 
+    mutate(rn = row_number()) %>% 
+    filter(rn == 1) %>% 
+    select(country, gdpPerCap)
+
+  return(outdata)
+
+}