Shiny App won't deploy but it runs locally - r

Question

My Shiny app is running locally but won't deploy. I've looked through similar questions and tried going through simple things which could be causing it e.g., wrong filepaths, setwd but that's not the case.
Here is some stuff from the logs:
Error in value[3L] :
Execution halted
dataTableOutput, renderDataTable
Error in value[3L] :
Shiny application exiting ...
The following object is masked from ‘package:graphics’:
dataTableOutput, renderDataTable
Here are some of the more relevant bits of the app.R file:
preds_with_mins <- read.csv("data/preds_with_mins.csv",
stringsAsFactors = FALSE)
filename <- paste0("data/all_players_final/all_players_finalgw", season,"_",current_gw, ".csv")
xg_data <- read.csv(filename, stringsAsFactors = FALSE) %>%
dplyr::rename('npxg90' = "npxG.90",
'xa90'= "xA.90") %>%
dplyr::select(player_name, npxg90, xa90, Team)
xg_data <- normal_preds %>%
dplyr::select(Player, Price, Position) %>%
dplyr::left_join(xg_data, by = c("Player" = "player_name"))
# xg_data[xg_data$Player == "Rúben Dias", "Price"] <- 6
# normal_preds[normal_preds$Player == "Rúben Dias", "Player"] <- "Ruben Dias"
# normal_preds[normal_preds$Player == "João Cancelo", "Player"] <- "Joao Cancelo"
review_data <- read.csv("data/fplreview_mins.csv",
stringsAsFactors = FALSE,
encoding = "UTF-8") %>%
dplyr::select(2,3,4,6,8,10,12,14,16,1) %>%
dplyr::rename('Price' = 'BV') %>%
dplyr::select(1,2, 5:9, 3:4,10) %>%
dplyr::rename('Position' = 'X.U.FEFF.Pos')
review_data[review_data$Position == "D", "Position"] <- "Defenders"
review_data[review_data$Position == "M", "Position"] <- "Midfielders"
review_data[review_data$Position == "F", "Position"] <- "Forwards"
review_data[review_data$Position == "G", "Position"] <- "Goalkeepers"
rev_gw_cols <- paste0("X", c(first_gw:(first_gw + 4)), "_Pts")
review_total <- review_data %>%
tidyr::pivot_longer(rev_gw_cols,names_to = "Gameweek",values_to = "points") %>%
dplyr::select(ID, Name,Team,Position,Price,Gameweek,points)%>%
dplyr::group_by(ID,Name,Team,Position,Price) %>%
dplyr::summarise(`Review Total`=round(sum(points),1)) %>%
dplyr::ungroup() %>%
dplyr::select(ID,
Team,
Name,
`Review Total`,
Price,
Position) %>%
rename('Price (£m)' = 'Price',
'Player' = 'Name') %>%
filter(`Review Total` > 0) %>%
dplyr::arrange(desc(`Review Total`)) %>%
dplyr::filter(Position != "Goalkeepers")
ae_total <- preds_with_mins %>%
dplyr::select(1,2,4,6,8,10, id) %>%
tidyr::pivot_longer(rev_gw_cols,names_to = "Gameweek",values_to = "points") %>%
dplyr::select(id, player_name,Gameweek,points) %>%
dplyr::group_by(id,player_name) %>%
dplyr::summarise(`AE Total`=round(sum(points),1)) %>%
dplyr::ungroup()
all_model_preds <- review_total %>%
dplyr::left_join(ae_total,
by = c("ID" = "id")) %>%
dplyr::filter(!is.na(`AE Total`)) %>%
dplyr::select(Player,
`AE Total`,
`Review Total`,
`Price (£m)`,
Position,
Team)
levels(all_model_preds$Position) <- c("Defenders","Midfielders","Forwards")
#ui / front end
ui <- fluidPage(
titlePanel(tags$h1("Albert's FPL Model" ,align="center")),
sidebarLayout(
sidebarPanel(#h3("Filters"),
sliderInput("gwrange","Gameweek Range",value=c(first_gw,first_gw + 3),min = first_gw, max = 38, step=1),
sliderInput("price_filter","Price (£m)",value=c(0,13.5),min=0,max=13.5,step=0.5),
selectInput("plot_pos",label="Position",choices=c("All Players","Defenders","Midfielders","Forwards")),
selectInput("team_filter",label="Team",choices=c("All Teams", sort(unique(normal_preds$Team)))),
selectInput("time_decay", label = "Time Decay", choices = c("Yes", "No"))
,width=2) #,textOutput("gwrange") #h3("Gameweek Range:3-6"),
,
mainPanel(
tabsetPanel(
tabPanel("Albert's Model",
DT::dataTableOutput("dynamic_df"),downloadButton('download','Download')
),
# tabPanel("Comparison With Review",
# DT::dataTableOutput("all_models")),
tabPanel("Points vs Price Graph",plotlyOutput("plot", width = "800px")
),
tabPanel("xG graphs",plotlyOutput("xgplot", width = "800px")
),
tabPanel("Methodology",htmlOutput("text"))
)#,width=10
)
)
)
#server/back end
server <- function(input, output, session) {
#summarise data based on the gw range they've chosen
gws_cols <- reactive(paste0("gw",seq(input$gwrange[1],input$gwrange[2],1)))
#gws_cols <- paste0("gw",gws())
predictions <- shiny::reactive(
if (input$time_decay == "No") {
normal_preds
} else {
decayed_preds
})
all_model_preds2 <- shiny::reactive(
if (input$time_decay == "No") {
all_model_preds
} else {
all_model_preds
})
base_data <- reactive(pivot_longer(predictions(),gws_cols(),names_to = "Gameweek",values_to = "Expected_points") %>%
select(Player,Team,Position,Price,Gameweek,Expected_points)%>%
group_by(Player,Team,Position,Price) %>%
summarise(Points=round(sum(Expected_points),1)) %>%
ungroup() %>%
dplyr::select(Team,
Player,
Points,
Price,
Position) %>%
rename('Price (£m)' = 'Price') %>%
filter(Points>0) %>%
arrange(desc(Points)) )
price_filtered_data <- reactive(base_data() %>%
dplyr::filter(`Price (£m)` >= input$price_filter[1],
`Price (£m)` <= input$price_filter[2]))
#output$static <- renderTable(head(all_preds))
#table_data <- reactive(if(input$plot_pos=="All Players"){filter(select(base_data(),-Position),`Price (£m)` >= input$price_filter[1] & `Price (£m)` <= input$price_filter[2])}else{ select(filter(base_data(),Position==input$plot_pos & `Price (£m)` >= input$price_filter[1] & `Price (£m)` <= input$price_filter[2], Team == input$team_filter),-Position)})
table_data <- reactive(if(input$plot_pos=="All Players" & input$team_filter == "All Teams"){
price_filtered_data()} else if (input$plot_pos=="All Players" & input$team_filter != "All Teams"){
dplyr::filter(price_filtered_data(),Team == input$team_filter)}
else if (input$plot_pos!="All Players" & input$team_filter == "All Teams"){
dplyr::filter(price_filtered_data(),Position==input$plot_pos)} else {
dplyr::filter(price_filtered_data(),Position==input$plot_pos & Team == input$team_filter)
})
output$dynamic_df <- DT::renderDataTable(select(table_data(), -Team),options = list(pageLength = 10))#,extensions="buttons",buttons=c('csv')
price_filtered_data_rev <- reactive(all_model_preds2() %>%
dplyr::filter(`Price (£m)` >= input$price_filter[1],
`Price (£m)` <= input$price_filter[2]))
#output$static <- renderTable(head(all_preds))
#table_data <- reactive(if(input$plot_pos=="All Players"){filter(select(base_data(),-Position),`Price (£m)` >= input$price_filter[1] & `Price (£m)` <= input$price_filter[2])}else{ select(filter(base_data(),Position==input$plot_pos & `Price (£m)` >= input$price_filter[1] & `Price (£m)` <= input$price_filter[2], Team == input$team_filter),-Position)})
table_data_rev <- reactive(if(input$plot_pos=="All Players" & input$team_filter == "All Teams"){
price_filtered_data_rev()} else if (input$plot_pos=="All Players" & input$team_filter != "All Teams"){
dplyr::filter(price_filtered_data_rev(),Team == input$team_filter)}
else if (input$plot_pos!="All Players" & input$team_filter == "All Teams"){
dplyr::filter(price_filtered_data_rev(),Position==input$plot_pos)} else {
dplyr::filter(price_filtered_data_rev(),Position==input$plot_pos & Team == input$team_filter)
})
#output$all_models <- DT::renderDataTable(table_data_rev(),options = list(pageLength = 10), rownames = FALSE)
output$download <- downloadHandler(filename=function(){"albertsfplmodel.csv"},
content=function(fname){
write.csv(preds_with_mins,fname)
})
#output$plot <- renderPlotly({if (input$plot_pos=="Defenders"){
# ggplotly(plot1, tooltip = "text") #ggplotly(plot1, tooltip = "text")
#} #else if (input$plot_pos=="Midfielders"){
# plot2
#
#} else {
# plot3
#}}) #, res = 96
plot_data <- reactive(if(input$plot_pos=="All Players" & input$team_filter == "All Teams"){
base_data()} else if (input$plot_pos=="All Players" & input$team_filter != "All Teams"){
dplyr::filter(base_data(),Team == input$team_filter)}
else if (input$plot_pos!="All Players" & input$team_filter == "All Teams"){
dplyr::filter(base_data(),Position==input$plot_pos)} else {
dplyr::filter(base_data(),Position==input$plot_pos & Team == input$team_filter)
})
# plot_cols <- reactive(if(input$plot_pos=="All Players"))
output$plot <- renderPlotly({ggplotly(ggplot(plot_data(),
aes(x=`Price (£m)`,y=Points,group=1,col=Position,
text=paste("Player: ",Player,"<br> Points: ",Points,"<br> Price: £",`Price (£m)`,"m",sep="")))+
geom_point() +
xlab("Price (£m)")+
ylab("Points")+scale_color_manual(values=c("Defenders"="#F8766D","Midfielders"="#00BA38","Forwards"="#619CFF")),tooltip="text")})
xg_plot_data <- reactive(if(input$plot_pos=="All Players" & input$team_filter == "All Teams"){
xg_data} else if (input$plot_pos=="All Players" & input$team_filter != "All Teams"){
dplyr::filter(xg_data,Team == input$team_filter)}
else if (input$plot_pos!="All Players" & input$team_filter == "All Teams"){
dplyr::filter(xg_data,Position==input$plot_pos)} else {
dplyr::filter(xg_data,Position==input$plot_pos & Team == input$team_filter)
})
output$xgplot <- renderPlotly({ggplotly(ggplot(xg_plot_data(),
aes(x=xa90,y=npxg90,group=1,col=Position,
text=paste("Player: ",Player,"<br> Price: £",Price,"m",sep="")))+
geom_point() +
xlab("xA/90")+
ylab("npxG/90")+scale_color_manual(values=c("Defenders"="#F8766D","Midfielders"="#00BA38","Forwards"="#619CFF")),tooltip="text")})
output$text <- renderText({paste0(h3("Data"),"The data used for the model is expected goal and assist data from www.Fbref.com " ,"</p>",h3("Goal and Assist Points"),"</p>",
"A weighted average of the last 20 games is used to calculate a player's average non-penalty expected goals per 90 (npxg/90) and expected assists per 90 (xA/90).
A player's average npxg/90 and xa/90 are adjusted for each game by the defensive strength of the opposition and if they're playing home or away. This is then multiplied by the number of points scored for a goal or assist.",h3("Clean Sheet Points"),"</p>",
"Weighted averages of non-penalty expected goals are also used to determine the attacking and defensive strength of each team.","</p>","To estimate points from clean sheets, a team's defensive strength is adjusted by the attacking strength of the opposition team and if they are playing at home or away. Their adjusted defsenive strength is then used as the mean of a poisson distribtution, to work out the probability of conceding zero goals.
This probability is then multiplied by the number of points scored for keeping a clean sheet.","</p>",h3("Miscallaneous"),"</p>","Players may also have their npxg/90 average modified depending on their finishing skill and if they take penalties.",
"</p>","All Predictions are adjusted by the number of minutes each player is estimated to play, with estimates taken from www.fplreview.com. This means players who are currently injured and more likely to be rotated are predicted fewer points.","</p>","Predictions for new signings or promoted players are not available or inaccurate due to lack of data.",
"</p>", "Alpha = 0.8 is used for the decayed predictions.")})
#output$gwrange <- renderText({"These Predictions are for gameweeks 3 to 6"})
#+
#scale_color_manual(values=c("#F8766D","00BA38","619CFF"))
#
#original mins_data df
#output$mins_data_original <- DT::renderDataTable(mins_data,options = list(pageLength = 10))
#reactive values to store mins_data df (which gets updated)
#values <- reactiveValues(df = mins_data)
#updating the df when user updates their estimates of a player's mins
#observeEvent(input$update_mins,{values$df[values$df$player_name==input$player_name,"avg_mins"] <- input$new_mins})
#output$mins_data_dynamic <- DT::renderDataTable(values$df)#,options = list(pageLength = 10))
}
shinyApp(ui, server)
It deployed fine before but I'm now having problems due to the 'all_model_preds' dataframe.
I've written this in a rush as I'm about to head out, please let me know if I need to be more specific