Почему строки моего фактора (Hour_of_Day) стали NA, когда я загрузил свое локальное блестящее приложение в shinyapps.io?
Недавно я создал приложение Shiny для отслеживания количества парковочных мест, выданных за день, с разбивкой по часам и с фильтрацией по срокам, местоположению, нарушению и должностному лицу. Приложение работает нормально, когда я запускаю его локально:
Правильный график вывода изображения
# Citations by Time of Day App
#Set / Read Directory
#-------------------------------------------------------------------------------#
#-------------------------------------------------------------------------------#
#Load Packages
library("tidyverse")
library("hms")
library("lubridate")
library("shiny")
library("shinyWidgets")
library("shinydashboard")
library("DT")
#Read Parking Dataset
x <- read.csv("Shiny Parking Trial Dataset.txt")
#Reformat Data
#Select and Relabel Variables
y <- x %>% select("Citation Number" = CON_TICKET_ID, "Officer" = STF_DESCRIPTION, "Location" = CLM_DESCRIPTION, "Violation" = VIC_DESCRIPTION)
#Update Officer Strings
updated_officer <- y %>% mutate_all(~str_replace_all(., "[//(//)]", ""))
updated_officer_2 <- updated_officer %>% mutate_all(~str_replace_all(., " Mobile Device User", ""))
updated_officer_3 <- updated_officer_2 %>% mutate_all(~str_replace_all(., " Mobile Device Use", ""))
violation_parse <- updated_officer_3 %>% mutate_all(~str_replace_all(., " -", ""))
# Create Variables for Dates and Times
set_dates <- as.data.frame(mdy_hms(x$CON_ISSUE_DATE))
Q <- set_dates %>% select("Issue_Time" = `mdy_hms(x$CON_ISSUE_DATE)`)
Hours <- format(Q, "%H%:%M:%S")
time_to_character <- as.character(Hours$Issue_Time)
Issue_Time <- parse_hms(time_to_character)
comb <- cbind(violation_parse,Issue_Time)
FINAL <- comb %>% mutate(Hour = hour(comb$Issue_Time))
FINAL_ADD_STRING <- as.character(FINAL$Hour)
ZZZ <- factor(FINAL_ADD_STRING, levels = c("0","1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24"), exclude = NA)
levels(ZZZ) <- list("12AM" = "0","1AM" = "1","2AM" = "2","3AM" = "3","4AM" = "4","5AM" = "5","6AM" = "6","7AM" = "7","8AM" = "8","9AM" = "9","10AM" = "10","11AM" = "11","12PM" = "12","1PM" = "13","2PM" = "14","3PM" = "15","4PM" = "16","5PM" = "17","6PM" = "18","7PM" = "19","8PM" = "20","9PM" = "21","10PM" = "22","11PM" = "23","12.AM" = "24")
Hour_of_Day <- ZZZ
List_Date_Time <- format(Q$Issue_Time, "%m/%d/%y %I:%M:%S %p")
Reformat_Date <- date(Q$Issue_Time)
#Put Everything Together
Clean_Complit <- cbind.data.frame(List_Date_Time,FINAL,Hour_of_Day,Reformat_Date)
Clean_Complete <- Clean_Complit %>% rename("Date / Time" = List_Date_Time)
# Create Lists for Violation, Location, and Officer
# List Violation
Violation_Count <- Clean_Complete %>% count(Violation)
Violation_Grab <- Violation_Count %>% select(Violation)
Violation_List <- as.list(Violation_Grab)
as.data.frame(Violation_Grab)
Violation_Options <- data.frame(Violation_Grab, row.names = Violation_Grab$Violation)
# List Location
Location_Count <- Clean_Complete %>% count(Location)
Location_Grab <- Location_Count %>% select(Location)
Location_List <- as.list(Location_Grab)
as.data.frame(Location_Grab)
Location_Options <- data.frame(Location_Grab, row.names = Location_Grab$Location)
# List Officer
Officer_Count <- Clean_Complete %>% count(Officer)
Officer_Grab <- Officer_Count %>% select(Officer)
Officer_List <- Officer_Grab
as.data.frame(Officer_Grab)
Officer_Options <- data.frame(Officer_Grab, row.names = Officer_Grab$Officer)
#Preset Colors
barfill <- "forestgreen"
barfill_2 <- "blue4"
barlines <- "black"
#Establish Theme
hw <- theme_gray()+ theme(
plot.title=element_text(hjust=0.5, size = 24),
axis.title.x = element_text(size=14),
axis.title.y = element_text(size=14),
plot.subtitle=element_text(hjust=0.5),
plot.caption=element_text(hjust=-.5),
# strip.text.y = element_blank(),
strip.background=element_rect(fill=rgb(.9,.95,1),
colour=gray(.5), size=.2),
panel.border=element_rect(fill=FALSE,colour=gray(.70)),
panel.grid.minor.y = element_blank(),
panel.grid.minor.x = element_blank(),
panel.spacing.x = unit(0.10,"cm"),
panel.spacing.y = unit(0.05,"cm"),
# axis.ticks.y= element_blank()
axis.ticks=element_blank(),
axis.text=element_text(colour="black"),
axis.text.x = element_text(size = 9),
axis.text.y=element_blank())
hy <- theme_gray()+ theme(
plot.title=element_text(hjust=0.5, size = 24),
axis.title.x = element_text(size=20),
axis.title.y = element_text(size=14),
plot.subtitle=element_text(hjust=0.5),
plot.caption=element_text(hjust=-.5),
# strip.text.y = element_blank(),
strip.background=element_rect(fill=rgb(.9,.95,1),
colour=gray(.5), size=.2),
panel.border=element_rect(fill=FALSE,colour=gray(.70)),
panel.grid.major.y = element_blank(),
panel.grid.minor.y = element_blank(),
panel.grid.minor.x = element_blank(),
panel.spacing.x = unit(0.10,"cm"),
panel.spacing.y = unit(0.05,"cm"),
# axis.ticks.y= element_blank()
axis.ticks=element_blank(),
axis.text=element_text(colour="black"),
axis.text.x = element_blank(),
axis.text.y = element_text(size = 12))
#-------------------------------------------------------------------------------#
#-------------------------------------------------------------------------------#
# Shiny Application
ui <- fluidPage(
titlePanel(h1("SP+ Citation Application", align = "right")),
sidebarLayout(
sidebarPanel(width = 3,
dateRangeInput("dates", label = "Date range", start = Sys.Date()-7, end = Sys.Date()-7, min = min(Reformat_Date), max = max(Reformat_Date)),
pickerInput("Location",label = "Location", choices = rownames(Location_Options),
multiple = TRUE, selected = rownames(Location_Options)[], options = list(`actions-box` = TRUE,`none-selected-text` = "No Locations", `selected-text-format` = "count > 3",
`count-selected-text` = "ALL", `live-search` = TRUE
)),
pickerInput('Violation', 'Violation', choices = rownames(Violation_Options),
multiple = TRUE, selected = rownames(Violation_Options)[],options = list(`actions-box` = TRUE,`none-selected-text` = "No Locations", `selected-text-format` = "count > 3",
`count-selected-text` = "ALL", `live-search` = TRUE
)),
pickerInput('Officer', 'Officer', choices = rownames(Officer_Options),
multiple = TRUE,selected = rownames(Officer_Options)[], options = list(`actions-box` = TRUE,`none-selected-text` = "No Locations", `selected-text-format` = "count > 3",
`count-selected-text` = "ALL", `live-search` = TRUE))
),
mainPanel(width = 9,
tabsetPanel(tabPanel("Time of Day",textOutput("TOTAL"), align = "right",plotOutput("TOD")),
tabPanel("Data Table", dataTableOutput("Table")),
#tabPanel("WOOP", dataTableOutput("Tester")),
tabPanel("Breakdown", radioGroupButtons(
inputId = "LVO_Filter",
label = "",
choices = c("Location","Violation","Officer"),justified = TRUE),
plotOutput("LocationBD", width = "100%", height = "700"))))
))
server <- function(input, output, session) {
#User subsets dataframe
NEW_df <- reactive({
m <- Clean_Complete %>% select(everything()) %>% filter(Clean_Complete$Reformat_Date >= input$dates[1] & Clean_Complete$Reformat_Date <= input$dates[2],Clean_Complete$Location %in% input$Location, Clean_Complete$Violation %in% input$Violation, Clean_Complete$Officer %in% input$Officer)
})
#Create Reactive Maximum Value for Y-Axis on Time of Day Plot
upper_y_value <- reactive({
Count_Hour_of_Day <- NEW_df() %>% count(Hour_of_Day)
Hour_of_Day_calc <- max(Count_Hour_of_Day$n) + sd(Count_Hour_of_Day$n / 4) + 5
print(Hour_of_Day_calc)
})
#Display Total Count of Citations
Total_Citations <- reactive({
numb <- nrow(NEW_df())
numb <- as.character(numb)
paste("Number of Parking Citations Issued by Hour (",numb,")")
})
#Produce Time of Day Plot
output$TOD <- renderPlot({
ggplot(NEW_df(), aes(x = Hour_of_Day)) +
geom_bar(color = barlines, fill = barfill, width = 1,position = position_nudge(x = 0.5)) +
scale_x_discrete(drop=F) +
geom_text(stat='count', aes(label=..count..),nudge_x = 0.5,vjust = -0.5, size = 5) +
scale_y_continuous(name = "Citations Issued",expand = c(0,0)) +
ggtitle("Number of Parking Citations Issued by Hour") +
xlab("Time of Day (Hour)") +
hw + expand_limits(y=c(0,upper_y_value()))
})
#Produce Data Table
output$Table <- renderDataTable({ NEW_df()[1:5]
})
#Location, Violation, Officer (LVO) Breakdown
spitout <- reactive({
LocCal <- NEW_df() %>% group_by_(input$LVO_Filter) %>% summarize(count = n()) %>% arrange(desc(count))
LocCal <- as.data.frame(LocCal)
if(nrow(LocCal) > 29){
subLocCal <- LocCal[(1:29),]
subLocCal <- as.data.frame(subLocCal)
} else {
subLocCal <- LocCal[]
subLocCal <- as.data.frame(subLocCal)
}
})
#Create Reactive Maximum Value for Y-Axis on Breakdown Plots
upper_y_value_breakdown <- reactive({ 1.25 * max(spitout()$count) })
#Function for Breakdown Plots
produce_breakdown_fx <- function(indata){
df <- indata
ggplot(df, aes(x = reorder(df[,1],df[,2]), y = df[,2])) +
geom_bar(color = barlines, fill = barfill_2, width = 1, stat = "identity") +
scale_x_discrete(drop=F) +
scale_y_continuous(name = "Number of Citations",expand = c(0,0)) +
geom_text(stat= 'identity', aes(label = df[,2]), size = 5, nudge_y = upper_y_value_breakdown() * 0.03) +
ggtitle(paste(" \n Count by", input$LVO_Filter)) +
xlab("") +
hy + expand_limits(y=c(0,upper_y_value_breakdown())) + coord_flip()
}
# Produce Breakdown Plots
output$LocationBD <- renderPlot({
produce_breakdown_fx(spitout())
})
#Citation Title
Total_Citations <- reactive({
numb <- nrow(NEW_df())
numb <- as.character(numb)
paste("(Total Count: ",numb,")")
})
output$TOTAL <- renderText({Total_Citations()})
}
# Run the application
shinyApp(ui = ui, server = server)
Когда я загрузил приложение в свою учетную запись shinyapps.io ( https://cypher-trial.shinyapps.io/Citation-Demo-App/), все работает нормально, за исключением того, что график времени дня отображает все цитаты, выданные как NA вдоль оси "Время дня (час)".
shinyapps.io Неверное изображение вывода графика
Журнал указывает, что проблема связана с фактором 'Hour_of_Day', содержащим неявный NA:
2019-02-26T23:43:21.998404+00:00 shinyapps[741215]: Warning: Factor `Hour_of_Day` contains implicit NA, consider using `forcats::fct_explicit_na`
2019-02-26T23:43:21.999019+00:00 shinyapps[741215]: [1] NA
Я пытался использовать рекомендованный forcats::fct_explicit_na решение, но мой график времени суток остался прежним. Я не уверен относительно того, что приводит к неправильной работе Hour_of_Day и приводит только к NA, особенно учитывая, что локальное приложение Shiny, кажется, работает отлично. Будем очень благодарны за любые предложения о том, как я могу решить проблему!
ОБНОВЛЕНИЕ: я решил проблему, подсказав "Фактор Hour_of_Day содержит неявное NA"Предупреждение, но очевидно, что предупреждение не имело никакого отношения к тому, что мой фактор" Hour_of_Day "стал NA для каждой строки. В журнале больше не отображаются какие-либо ошибки или предупреждения при запуске приложения, так что я еще больше растерялся, как чтобы исправить мою проблему.
РЕШЕНО: Проблема обнаружена, и приложение теперь исправлено! Мой код имел дополнительный "%" после "H" в следующем утверждении формата.
Hours <- format(Q, "%H%:%M:%S")