R Shiny Highcharter - Add series without full reload

Question

I'm plotting a financial time serie on an R-Shiny project using the highcharter package and I want to allow users to add indicators (such as moving averages) on the chart. In order to do that, I conditioned an hc_add_series_xts() in an if statement server-side. It works fine, but everytime someone adds an indicator the chart fully reload while I'd prefer a dynamic update. Is there a way to make it?

To better explain my case, It follows a single page Shiny example:

library(shiny)
library(quantmod)
library(highcharter)
library(TTR)


ui <- fluidPage(

      mainPanel(
        highchartOutput("hc"),
        checkboxInput("indicator","Add Moving Average")
      )
)


server <- function(input, output) {

  output$hc <- renderHighchart({

    y <- getSymbols("MSFT", auto.assign=FALSE)

    hc <- highchart() %>%
          hc_exporting(enabled = TRUE)%>%
          hc_add_series_ohlc(y) %>% 
          hc_add_theme(hc_theme_538(colors = c("red", "blue", "green"),
                                    chart = list(backgroundColor = "white") ))

    if(input$indicator){hc <- hc %>% hc_add_series_xts(name = "Moving Average", EMA(Cl(y)))}

    return(hc)
  })

}


shinyApp(ui = ui, server = server)

PS: I intentionally call all historical prices avaible in order to make perceptible the "full reaload" event.

Answer 1

Currently is not posible update the chart using shiny beacuse, as you see, chart a new chart.

Have you tried hc_legend(enabled = TRUE) ? So you can show or hide a series clicking over the legend.

hc <- highchart() %>%
  hc_exporting(enabled = TRUE)%>%
  hc_add_series_ohlc(y) %>% 
  hc_add_theme(hc_theme_538(colors = c("red", "blue", "green"),
                            chart = list(backgroundColor = "white") )) %>% 
  hc_add_series_xts(name = "Moving Average", EMA(Cl(y)), visible = FALSE) %>% 
  hc_legend(enabled = TRUE)

hc

Hope this help

Answer 2

With your example, jbkunst method is much better than mine. Especially because the rendering of the graph (in ui ) takes much longer than building it (in server ).

However in some cases, the solution of loading your basic object in a variable of the server function, then add only new indicators when they are selected can improve latency:

server <- function(input, output) {
  y <- getSymbols("MSFT", auto.assign=FALSE)
  hc <- highchart() %>%
          hc_exporting(enabled = TRUE)%>%
          hc_add_series_ohlc(y) %>% 
          hc_add_theme(hc_theme_538(colors = c("red", "blue", "green"),
                                    chart = list(backgroundColor = "white") ))

  output$hc <- renderHighchart({
    hc_now <- hc
    if(input$indicator){hc_now <- hc_now %>% hc_add_series_xts(name = "Moving Average", EMA(Cl(y)))}
    return(hc_now)
  })
}

The first time it takes a while to load, but changing indicators takes less time