I'm trying to publish a shiny application on connect.
My application has the following script:
library(shiny)
library(googleCloudStorageR)
library(gargle)
if(!require('readr')) {
install.packages('readr')
library('readr')
}
if(!require('dplyr')) {
install.packages('dplyr')
library('dplyr')
}
if(!require('dygraphs')) {
install.packages('dygraphs')
library('dygraphs')
}
if(!require('plotly')) {
install.packages('plotly')
library('plotly')
}
if(!require('PerformanceAnalytics')) {
install.packages('PerformanceAnalytics')
library('PerformanceAnalytics')
}
if(!require('shinydashboard')) {
install.packages('shinydashboard')
library('shinydashboard')
}
if(!require('xts')) {
install.packages('xts')
library('xts')
}
if(!require('zoo')) {
install.packages('zoo')
library('zoo')
}
if(!require('rjson')) {
install.packages('rjson')
library('rjson')
}
if(!require('googleAuthR')) {
install.packages('googleAuthR')
library('googleAuthR')
}
scope <-c("https://www.googleapis.com/auth/cloud-platform")
token <- token_fetch(scopes = scope)
gcs_auth(token = token))
p1 <- gcs_get_object("p1.csv", bucket = "my_bucket")
p2 <- gcs_get_object("p2.csv", bucket = "my_bucket")
p3 <- gcs_get_object("p3.csv", bucket = "my_bucket")
portfolio_choices <- c(
"Conservative" = "conservative_portfolio_returns",
"Balanced" = "balanced_portfolio_returns",
"Aggressive" = "aggressive_portfolio_returns"
)
ui <- dashboardPage(
dashboardHeader(title = "Portfolio Dashboard"),
dashboardSidebar(
selectInput(
"portfolio",
"Choose a portfolio",
choices = portfolio_choices,
selected = "balanced_portfolio_returns"
),
dateInput(
inputId = "date",
label = "Starting Date",
value = "2010-01-01",
format = "yyyy-mm-dd"
),
sliderInput("mar", "Min Acceptable Rate", min = 0, max = 0.1, value = 0.008, step = 0.001),
numericInput("window", "Rolling Window", min = 6, max = 36, value = 12)
),
dashboardBody(
fluidRow(
box(title = "Rolling Sortino", width = 12,
plotlyOutput("time_series")
)
),
fluidRow(
box(title = "Scatterplot", width = 4,
plotlyOutput("scatterplot", height = 250)
),
box(title = "Histogram", width = 4,
plotlyOutput("histogram", height = 250)
),
box(title = "Density", width = 4,
plotlyOutput("density", height = 250)
)
)
)
)
server <- function(input, output) {
rate_limit_sec <- 2
portfolio_selected <- throttle(reactive({
req(input$portfolio, input$date)
if (input$portfolio =="conservative_portfolio_returns") {
returns <- p1}
else if(input$portfolio=="balanced_portfolio_returns"){
returns <- p2}
else {
returns <- p3}
returns %>%
as_tibble() %>%
#collect() %>%
mutate(date = as.Date(date)) %>%
filter(date >= input$date)
}), rate_limit_sec * 1000)
rolling_sortino <- reactive({
req(input$mar)
req(input$window)
portfolio_selected()$returns %>%
xts::xts(order.by = portfolio_selected()$date) %>%
rollapply(input$window, function(x) SortinoRatio(x, MAR = input$mar)) %>%
`colnames<-`("24-rolling")
})
sortino_byhand <- reactive({
portfolio_selected() %>%
mutate(ratio = mean(returns - input$mar) / sqrt(sum(pmin(returns - input$mar, 0)^2) / nrow(.))) %>%
# Add two new columns to help with ggplot.
mutate(status = ifelse(returns < input$mar, "down", "up"))
})
output$time_series <- renderPlotly({
plot_ly() %>%
add_lines(x = index(rolling_sortino()), y = as.numeric(rolling_sortino())) %>%
layout(
hovermode = "x",
xaxis = list(
rangeslider = list(visible = TRUE),
rangeselector = list(
x = 0, y = 1, xanchor = 'left', yanchor = "top", font = list(size = 9),
buttons = list(
list(count = 1, label = 'RESET', step = 'all'),
list(count = 1, label = '1 YR', step = 'year', stepmode = 'backward'),
list(count = 3, label = '3 MO', step = 'month', stepmode = 'backward'),
list(count = 1, label = '1 MO', step = 'month', stepmode = 'backward')
)
)
)
)
})
output$scatterplot <- renderPlotly({
portfolio_scatter <- ggplot(sortino_byhand(), aes(x = date, y = returns, color = status) )+
geom_point() +
geom_vline(xintercept = as.numeric(as.Date("2016-11-30")), color = "blue") +
geom_hline(yintercept = input$mar, color = "purple", linetype = "dotted") +
scale_color_manual(values = c("tomato", "chartreuse3")) +
theme(legend.position = "none") + ylab("percent monthly returns")
ggplotly(portfolio_scatter) %>%
add_annotations(
text = "Trump", x = as.numeric(as.Date("2016-11-30")),
y = -.05, xshift = -10, textangle = -90, showarrow = FALSE
)
})
output$histogram <- renderPlotly({
p <- ggplot(sortino_byhand(), aes(x = returns)) +
geom_histogram(alpha = 0.25, binwidth = .01, fill = "cornflowerblue") +
geom_vline(xintercept = input$mar, color = "green")
ggplotly(p) %>%
add_annotations(text = "MAR", x = input$mar, y = 10, xshift = 10, showarrow = FALSE, textangle = -90)
})
output$density <- renderPlotly({
sortino_density_plot <- ggplot(sortino_byhand(), aes(x = returns)) +
stat_density(geom = "line", size = 1, color = "cornflowerblue")
shaded_area_data <- ggplot_build(sortino_density_plot)$data[[1]] %>%
filter(x < input$mar)
sortino_density_plot <-
sortino_density_plot +
geom_area(data = shaded_area_data, aes(x = x, y = y), fill = "pink", alpha = 0.5) +
geom_segment(
data = shaded_area_data, aes(x = input$mar, y = 0, xend = input$mar, yend = y),
color = "red", linetype = "dotted"
)
ggplotly(sortino_density_plot) %>%
add_annotations(
x = input$mar, y = 5, text = paste("MAR =", input$mar, sep = ""), textangle = -90
) %>%
add_annotations(
x = (input$mar - .02), y = .1, text = "Downside",
xshift = -20, yshift = 10, showarrow = FALSE
)
})
}
shinyApp(ui = ui, server = server)
I don't understand because I have installed and loaded the "readr" package in my script.