Data object not found when deploying shiny app

Question

I am working on my first shiny app and am running into an issue where the data that is used to render my data table is not being picked up by shinyapps.io.

The app runs fine in my console, but when I deploy it the app opens in a browser with the error: object 'Pitchers' not found, where 'Pitchers' is one of my data objects.

I found one suggestion that said to place your data in a folder within the app's folder, but that still is not working.

Here is my current server.R code:

shinyServer(function(input, output) {

  Pitchers <- read.csv("data/Pitchers_Edge.csv", header=TRUE, check.names = FALSE)
  Batters <- read.csv("data/Batters_Edge.csv", header=TRUE, check.names = FALSE)

  output$table1 <- renderDataTable({

    if (input$Year != "All"){
      Pitchers <- Pitchers[Pitchers$Year == input$Year,]
    } 
    Pitchers 
  })

  output$table2 <- renderDataTable({

    if (input$Year != "All"){
      Batters <- Batters[Batters$Year == input$Year,]
    } 
    Batters
  })

})

And here is the ui.R code:

shinyUI(fluidPage(
  titlePanel('Edge%: 2010-Present'),
    fluidRow(
      column(12,

                    p("Provides current and historical data on the percentage of pitches thrown to different parts of the strike zone by pitchers and to batters"),
                    p("Created and maintained by Bill Petti",
                      a("(@BillPetti)", href = "https://twitter.com/billpetti")),
                    p("Data last updated",Sys.time()))
      ),
  fluidRow(
    column(5, 
           selectInput("Year", 
                       "Year:", 
                       c("All", 
                         unique(as.character(Pitchers$Year)))))
    ),
  mainPanel(
      tabsetPanel(
        tabPanel("Pitchers", dataTableOutput(outputId = 'table1')),
        tabPanel("Batters", dataTableOutput(outputId = 'table2')),
        tabPanel("About", 
             br(), h1("About Edge%"), 
             br(), p("A few years ago, Jeff Zimmerman and I created a metric to represent how often a pitcher threw to the edges of the strike zone compared to the heart of the strike zone. The result was Edge%. The metric has evolved to include separate metrics for different edges (upper, lower, etc.). In the image below, the brown shaded areas represent the horizontal edges of the strike zone, the blue the top, and the green the bottom edges. You will notice the horizontal edges differ by batter handedness, given how umpires actually call balls and strikes."), 
             br(), img( src = "edge_image.png", height = 350, width = 700)), 
            br(), p("Edge% is useful in a number of contexts. For example, we know that as pitchers age they lose velocity and therefore need to avoid throwing to the heart of the plate to be successful. Edge% provides a quick look at who is adjusting to lower velocity and who isn't. It can also be used to see how pitchers are adjusting to hitters as they age (i.e. as hitters improve, pitchers may avoid the heart of the plate more, or as hitters decline they may begin challenge them more."), 
            br(), p("For more information on Edge%, check out these articles:", 
            br(), br(), a("Introduction to Edge%", href = "http://www.fangraphs.com/blogs/the-difference-pitching-on-the-edge-makes/"), 
            br(), br(), a("Collection of Articles Using and Expanding on Edge%", href = "http://www.fangraphs.com/blogs/category/edge/"),
            br(), br(), a("Most Recent Version", href = "http://www.hardballtimes.com/expanding-the-edges-of-the-strike-zone/") 
                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   )
                 )
)
)
)

I am deploying from a Windows PC.

Would appreciate any suggestions!

Update

I took Colin's advice and updated my code, but I am still running into the same issue.

Here's what my app should look like when deployed:

https://i.stack.imgur.com/idmvC.jpg

And here's what happens when I publish and view, either locally or on the web:

https://i.stack.imgur.com/ikBrD.jpg

It isn't just that the object is missing, but I can see the rest of the app. The entire app just disappears.

Here is my updated code:

library(shiny)

shinyServer(function(input, output) {

  Pitchers <- reactive({read.csv("data/Pitchers_Edge.csv", header=TRUE, check.names = FALSE)})
  Batters <- reactive({read.csv("data/Batters_Edge.csv", header=TRUE, check.names = FALSE)})

  output$table1 <- renderDataTable({
    Pitchers <- Pitchers()
    if (input$Year != "All"){
      Pitchers <- Pitchers[Pitchers$Year == input$Year,]
    } 
    subset(Pitchers, Pitchers$'# of total pitches'>= input$pitch_total) 
  })

  output$table2 <- renderDataTable({
    Batters <- Batters()
    if (input$Year != "All"){
      Batters <- Batters[Batters$Year == input$Year,]
    } 
    subset(Batters, Batters$'# of total pitches'>= input$pitch_total)
  })

})

shinyUI(fluidPage(
  titlePanel('Edge%: 2010-Present'),
  fluidRow(
    column(12,
           p("Provides current and historical data on the percentage of pitches thrown to different parts of the strike zone by pitchers and to batters"),
           p("Created and maintained by Bill Petti",
             a("(@BillPetti)", href = "https://twitter.com/billpetti")),
           p("Data last updated",Sys.time()))
  ), 
    sidebarLayout(
    sidebarPanel(selectInput("Year", 
                                    "Year:", 
                                    c("All", 
                                      unique(as.character(Pitchers$Year)))),
                 numericInput("pitch_total", 
                             label = "Minimum # of Pitches:",
                             value = 300)
  )
  ,
  mainPanel(
    tabsetPanel(
      tabPanel("Pitchers", dataTableOutput(outputId = 'table1')),
      tabPanel("Batters", dataTableOutput(outputId = 'table2')),
      tabPanel("About", 
               br(), h1("About Edge%"), 
               br(), p("A few years ago, Jeff Zimmerman and I created a metric to represent how often a pitcher threw to the edges of the strike zone compared to the heart of the strike zone. The result was Edge%. The metric has evolved to include separate metrics for different edges (upper, lower, etc.). In the image below, the brown shaded areas represent the horizontal edges of the strike zone, the blue the top, and the green the bottom edges. You will notice the horizontal edges differ by batter handedness, given how umpires actually call balls and strikes."), 
               br(), img( src = "edge_image.png", height = 350, width = 700), 
      br(), p("Edge% is useful in a number of contexts. For example, we know that as pitchers age they lose velocity and therefore need to avoid throwing to the heart of the plate to be successful. Edge% provides a quick look at who is adjusting to lower velocity and who isn't. It can also be used to see how pitchers are adjusting to hitters as they age (i.e. as hitters improve, pitchers may avoid the heart of the plate more, or as hitters decline they may begin challenge them more."), 
      br(), p("For more information on Edge%, check out these articles:"), 
              br(), a("Introduction to Edge%", href = "http://www.fangraphs.com/blogs/the-difference-pitching-on-the-edge-makes/"), 
              br(), br(), a("Collection of Articles Using and Expanding on Edge%", href = "http://www.fangraphs.com/blogs/category/edge/"),
              br(), br(), a("Most Recent Version", href = "http://www.hardballtimes.com/expanding-the-edges-of-the-strike-zone/") 
      )
    )))))

Answer 1

Oh! I think it's that you're not loading the pitchers file as part of the ui.R script. So when you're defining the year input it's not able to find Pitchers$year. Can you try reading the pitchers file in the ui script above the shinyUI() call?

Some stuff about data injest in shiny: If your data's not going to change regularly, you don't need your data-ingest to be reactive. Just put it at the very beginning of the server file before the shinyServer() call. That part of the code gets run just once, when you deploy the app so it's a good place to do any library() calls, data injest, or pre-processing stuff that doesn't depend upon the user in any way. You can just use read.csv() in the same was as you would in regular R, it's also sometimes good to save a binay of the dataframe because read.csv can be a bit slow.

If it is going to change (like an updated csv is going to be put in the file regularly) then put the read.csv() call below the shinyServer() call but again shouldn't need to be reactive.

Hope that helps!

Answer 2

Once you get inside the renderDataTable function, it doesn't recognize what Pitchers is because it was defined outside the function. However, if you set Pitchers to be a reactive element, it can be called within the function. So would set the variable and then call it as a function inside the render function.

Pitchers<- reactive({read.csv("Your Stuff")})

output$table1 <- renderDataTable(
Pitch<-Pitchers()

"Then call all your if statements and such on the function using Pitch"
)

Hope that helps!

Answer 3

Piggybacking on a previous answer:

I agree that the problem is that the object "Pitchers" needs to be defined in both shinyServer and shinyUI. Creating a third file called "global.R" may be the best way to solve this problem, since that way you only need to read the database one time (which is faster and nicer to the free shinyapps.io server). See https://shiny.rstudio.com/articles/scoping.html for more information about creating variables with a global scope.

Answer 4

I suggest that you must include library DT in UI and server functions, I solved it like :

library(DT)

shinyUI(....)

also before shinyServer(...)

I had the same problem. I hope this information is helpful. :)