Capture user input

[krose]

Nice work on the package!

This is a feature request. It would be great if I could capture and save the user input from shinydrawr. I don't know much about JavaScript so I have no clue about the implementation of it, but it's a feature that would be really useful to me, as I could benchmark user assumptions with real data.

[nstrayer]

It's your lucky day! This is already implemented. Just copy the following into a script and run run it and you should have a working demo. Currently it returns the drawn y-values corresponding the x-locations in your supplied data post your cutoff. Is this what you're looking for?

#
# A small demo app for the shinydrawr function
#
# devtools::install_github("nstrayer/shinysense")
library(shiny)
library(shinythemes)
library(shinysense)
library(tidyverse)


ui <- fluidPage(
  theme = shinytheme("flatly"),
  titlePanel("shinydrawr demo"),
  p("Try and guess the rest of the chart below"),
  hr(),
  fluidRow(
    column(width = 8,
           shinydrawrUI("outbreak_stats")),
    column(width = 3, offset = 1,
           h2("Drawn values:"), tableOutput("displayDrawn"))
  )
)


server <- function(input, output) {
  random_data <- data_frame(time = 1:30,
                            metric = time * sin(time / 6) + rnorm(30))

  #server side call of the drawr module
  drawChart <- callModule(
    shinydrawr,
    "outbreak_stats",
    data = random_data,
    draw_start = 15,
    x_key = "time",
    y_key = "metric",
    y_max = 20,
    y_min = -50
  )

  #logic for what happens after a user has drawn their values. Note this will fire on editing again too.
  observeEvent(drawChart(), {
    drawnValues = drawChart()

    drawn_data <- random_data %>%
      filter(time >= 15) %>%
      mutate(drawn = drawnValues)

    output$displayDrawn <- renderTable(drawn_data)
  })

}

# Run the application
shinyApp(ui = ui, server = server)

[krose]

Almost there :) Let's say I wan't to plot the data as points and then get the user to draw the best fit, but with points as well. Is that possible?

I have two use cases:

• If the user could add best fit points/line on the plot below.
• The user could also just get an x value and then add their y value as a single point on the canvas (not for the whole range).

with(do.call(rbind, lapply(1:10,  function(x){data.frame(x = 1:10, y = 1:10 + rnorm(10, sd = 2))})), plot(x, y))

[nstrayer]

Ah I see. That is a good idea and one I've been meaning to get around to. Tagging this with an enhancement tag and will hopefully get around to it soon enough.