Problems using gcIntermediate within R Shiny?

Question

I have been working on an app using R Shiny and everything works except one small part where I call gcIntermediate within the Shiny server. I was wondering if this is a typical issue or if I am just doing something horribly wrong which I am sadly oblivious to. I have a set of about 30'000 cities, their countries, coordinates, postal codes, etc. In part of this app, I would like to R to plot the great circles that connect points A and B on a world map. However, I have been unsuccessful due to an undefined value for draw_line when run within the Shiny environment. It does work just fine if I feed it values and run it outside of Shiny though. Hence, I'm stuck.

Here's the UI

#GC Plot Test UI

shinyUI(pageWithSidebar(

  headerPanel("Great Circles on a Map"),

  sidebarPanel(

    h3("Select Origin Codes"),

    uiOutput("choose_origin_code1"),

    uiOutput("choose_origin_code2"),

    h3("Select Destination Codes"),

    uiOutput("choose_destination_code1"),

    uiOutput("choose_destination_code2"),

    br()
  ),


  mainPanel(
    #tableOutput("MapPlot")

    plotOutput("MapPlot")

  )
))

Here's the server:

#GC Plot Test Server

library(shiny)
library(RColorBrewer)
library(ggplot2)
library(maps)
library(mapdata)
library(geosphere)
library(sp)

#Some random cities and locations from my larger data set
city_name <- c("Tunceli", "Udomlya", "Moscow", "Kaunas")
country_code <- c("TR", "RU", "RU", "LT")
city_code <- c(62, 17, 10, 44)
latitude <- c(39.108, 57.879, 55.752, 54.90)
longitude <- c(39.547, 34.993, 37.616, 23.90)
df <- data.frame(city_name, country_code, city_code, latitude, longitude)

map.dat <- map_data("world")

shinyServer(function(input, output) {

  output$choose_origin_code1 <- renderUI({
    selectInput("origin.code1", "Country of Origin", unique(as.character(df$country_code)))
  })

  output$choose_origin_code2 <- renderUI({
    #If missing input, return to avoid error later in function
    if(is.null(input$origin.code1))
      return()

    origin.code2.list1 <- df[grep(as.character(input$origin.code1), as.character(df$country_code)),]
    origin.code2.list2 <- unique(as.character(origin.code2.list1$city_code))

    #Create the checkboxes and deselect them all by default
    selectInput("origin.code2", "Source Location Code", origin.code2.list2)

  })


  output$choose_destination_code1 <- renderUI({
    #Check ship.from.code1 and ship.from.code2 input value
    if(is.null(input$origin.code1) | is.null(input$origin.code2))
      return()

    selectInput("dest.code1", "Destination Country", unique(as.character(df$country_code)))   

  })

  #Destination Location Input
  output$choose_destination_code2 <- renderUI({
    #Check ship.from.code1 and ship.from.code2 input value
    if(is.null(input$dest.code1))
      return()

    dest.code2.list1 <- df[grep(as.character(input$dest.code1), as.character(df$country_code)),]
    dest.code2.list2 <- unique(as.character(dest.code2.list1$city_code))

    selectInput("dest.code2", "Destination Location Code 2", dest.code2.list2)

  })

  #Map the routes
  output$MapPlot <- renderPlot({
    # If missing input, return to avoid error later in function
    if(is.null(input$origin.code1) | is.null(input$origin.code1) | is.null(input$dest.code1) | is.null(input$dest.code2))
      return()

    #Find the row with starting point information
    Origin_Row <- which(df$country_code == input$origin.code1 
                        & df$city_code == input$origin.code2)

    #Find the row with the destination information
    Dest_Row <- which(df$country_code == input$dest.code1
                      & df$city_code == input$dest.code2)

    #Make coordinate data for origin and destination explicit
    Origin_lat <- df$latitude[Origin_Row]
    Origin_lon <- df$longitude[Origin_Row]
    Dest_lat <- df$latitude[Dest_Row]
    Dest_lon <- df$longitude[Dest_Row]

    coords <- data.frame( c(Origin_lon, Dest_lon), c(Origin_lat, Dest_lat))
    colnames(coords) <- c("Lon", "Lat")

    draw_line <- gcIntermediate(coords[1,], coords[2,], n = 100, addStartEnd = TRUE)

    Map_center <- c(mean(coords[(1:2),1]), mean(coords[(1:2),2]), 0) #Centered on Zurich, CH

    p <- ggplot() 
    p <- p + geom_polygon(aes(long,lat, group=group), fill="grey65", data = map.dat) + theme_bw() + theme(axis.text = element_blank(), axis.title = element_blank()) 
    p <- p + geom_line(aes(x = draw_line[,1],
                           y = draw_line[,2]),
                       color = 'red')
    p <- p + coord_map("ortho", orientation = c(Map_center))
    p

  })

})

The error always occurs in the draw_line section where I call gcIntermediate(). I have attempted a number of different things for troubleshooting, but all I have come up with is that the gcIntermediate() function is not working in conjunction with Shiny. If this is the case, does anyone know how to fix it or of a possible work around?

Bonus: I'd like my plot to zoom in on the route when it is displayed, but I get distortion of the map when I do that, so I stuck to the world view. Does anyone know how to acheive that?

Answer 1

I found the issue was not putting the output of gcIntermediate into a data frame and redefining the data in the geom_line() layer.

 draw_line <- data.frame(gcIntermediate(coords[1,], coords[2,], n = 100, addStartEnd = TRUE))

      p <- ggplot() 
      p <- p + geom_polygon(aes(long,lat, group=group), fill="grey65", data = map.dat) + theme_bw() + theme(axis.text = element_blank(), axis.title = element_blank()) 
      p <- p + geom_line(data = draw_line, aes(x = draw_line[,1], y = draw_line[,2], color = 'red'))
      p <- p + coord_map("ortho", orientation = c(Map_center))
      p