action-graphics.Rmd

# Graphics

### Exercise 7.6.1 {-}

Make a plot with click handle that shows all the data returned in the input.

:::solution
#### Solution {-}

We can use the `allRows` argument in `nearPoints` to see the entire data and
add a boolean column that will be true `TRUE` for the given point (i.e., row)
that was clicked.

```{r, eval=FALSE}
library(shiny)
library(ggplot2)

ui <- fluidPage(
  plotOutput("plot", click = "plot_click"),
  tableOutput("data")
)

server <- function(input, output, session) {
  output$plot <- renderPlot({
    ggplot(mtcars, aes(wt, mpg)) + geom_point()
  }, res = 96)
  
  output$data <- renderTable({
    nearPoints(mtcars, input$plot_click, allRows = TRUE)
  })
}

shinyApp(ui, server)
```
:::

<!---------------------------------------------------------------------------->
<!---------------------------------------------------------------------------->
<!---------------------------------------------------------------------------->

### Exercise 7.6.2 {-}

Make a plot with click, dblclick, hover, and brush output handlers and nicely
display the current selection in the sidebar. Plot the plot in the main panel.

:::solution
#### Solution {-}

We can use the `nearPoints` function to extract the data from `plot_click`,
`plot_dbl`, and `plot_hover`. We need to use the function `brushedPoints` to
extract the points within the `plot_brush` area.

To 'nicely' display the current selection, we will use `dataTableOutput`.

```{r, eval=FALSE}
library(shiny)
library(ggplot2)

# Set options for rendering DataTables.
options <- list(
  autoWidth = FALSE,
  searching = FALSE,
  ordering = FALSE,
  lengthChange = FALSE,
  lengthMenu = FALSE,
  pageLength = 5, # Only show 5 rows per page.
  paging = TRUE, # Enable pagination. Must be set for pageLength to work.
  info = FALSE
)

ui <- fluidPage(
  
  sidebarLayout(
    sidebarPanel(
      width = 6,
      
      h4("Selected Points"),
      dataTableOutput("click"), br(),

      h4("Double Clicked Points"),
      dataTableOutput("dbl"), br(),
      
      h4("Hovered Points"),
      dataTableOutput("hover"), br(),
      
      h4("Brushed Points"),
      dataTableOutput("brush")
    ),
    
    mainPanel(width = 6,
              plotOutput("plot",
                         click = "plot_click",
                         dblclick = "plot_dbl",
                         hover = "plot_hover",
                         brush = "plot_brush")
    )
  )
)

server <- function(input, output, session) {
  
  output$plot <- renderPlot({
    ggplot(iris, aes(Sepal.Length, Sepal.Width)) + geom_point()
  }, res = 96)
  
  output$click <- renderDataTable(
    nearPoints(iris, input$plot_click),
    options = options)
  
  output$hover <- renderDataTable(
    nearPoints(iris, input$plot_hover),
    options = options)
  
  output$dbl <- renderDataTable(
    nearPoints(iris, input$plot_dbl),
    options = options)
  
  output$brush <- renderDataTable(
    brushedPoints(iris, input$plot_brush),
    options = options)
}

shinyApp(ui, server)
```
:::

<!---------------------------------------------------------------------------->
<!---------------------------------------------------------------------------->
<!---------------------------------------------------------------------------->

### Exercise 7.6.3 {-}

Compute the limits of the distance scale using the size of the plot.

```{r, eval=FALSE}
output_size <- function(id) {
  reactive(c(
    session$clientData[[paste0("output_", id, "_width")]],
    session$clientData[[paste0("output_", id, "_height")]]
  ))
}
```

:::solution
#### Solution {-}

Let us use the plot's width and height to estimate the scale limits for our
plot.

To verify that the recommended limits are correct, click around the plot
and watch how the distance scale changes on the legend. These values should
oscillate between the recommended limits.

Resize the browser's window to change the width and height reactives.

```{r, eval=FALSE}
library(shiny)
library(ggplot2)

df <- data.frame(x = rnorm(100), y = rnorm(100))

ui <- fluidPage(
  plotOutput("plot", click = "plot_click"),
  textOutput("width"),
  textOutput("height"),
  textOutput("scale")
)

server <- function(input, output, session) {
  
  # Save the plot's widht and height.
  width <- reactive(session$clientData[["output_plot_width"]])
  height <- reactive(session$clientData[["output_plot_height"]])
  
  # Print the plot's width, the plot's height, and the suggested scale limits.
  output$width <- renderText(paste0("Plot's width: ", width()))
  output$height <- renderText(paste0("Plot's height: ", height()))
  output$scale <- renderText({
    paste0("Recommended limits: (0, ", max(height(), width()), ")")
  })
  
  # Store the distance computed by the click event.
  dist <- reactiveVal(rep(1, nrow(df)))
  
  # Update the dist reactive as needed.
  observeEvent(input$plot_click, {
    req(input$plot_click)
    dist(nearPoints(df, input$plot_click, allRows = TRUE, addDist = TRUE)$dist_)
  })
  
  output$plot <- renderPlot({
    df$dist <- dist()
    ggplot(df, aes(x, y, size = dist)) +
      geom_point()
  })
}

shinyApp(ui, server)

```
:::