I have a dataset (tst_geo.csv) of the form:
lat, lon, time
10, 20, 1
10, 20, 2
10, 20, 3
40, 40, 4
40, 40, 5
40, 40, 6
0, 0, 7
0, 0, 8
0, 0, 9
R code:
library(shiny)
library(leaflet)
library(plyr)
ui <- fluidPage(
sidebarLayout(
sidebarPanel(
uiOutput("slider")
),
mainPanel(
leafletOutput("map")
)
)
)
server <- function(input, output, session){
df <- read.csv("tst_geo.csv", header=TRUE)
df['time'] <- as.numeric(df$time)
#make dynamic slider
output$slider <- renderUI({
sliderInput("time_span", "Time Span", step=1, min=min(df$time),
max=max(df$time), value = c(min(df$time), max(df$time)))
})
filter_df <- reactive({
df[df$time >= input$time_span[1] & df$time <= input$time_span[2], ]
})
output$map <- renderLeaflet(
leaflet() %>% addTiles()
)
observe({
points_df <- ddply(filter_df(), c("lat", "lon"), summarise, count = length(timestamp))
cat(nrow(points_df))
leafletProxy("map", data = points_df) %>% clearShapes() %>% addCircles()
})
}
shinyApp(ui, server)
I have a slider to only show points within a certain time range.
However, inside the observe
function the points are not being cleared when I call clearShapes()
.
Do you have any ideas why this is happening?
The culprit is renderUI()
in this case. Because you used renderUI()
, the rendering of the slider is delayed. By the time the observer runs for the first time, the slider is not there yet, and input$time_span
is initially NULL
, so filter_df()
returns an empty data frame. I don't see a particular reason of using renderUI()
in this case (perhaps you have a reason), and you can either just move sliderInput()
to ui.R, or check if (is.null(input$time_span))
before you add circles to the map, or change observe()
to observeEvent()
(if you are using a recent version of shiny):
observeEvent(input$time_span, {
points_df <- ddply(filter_df(), c("lat", "lon"), summarise, count = length(timestamp))
cat(nrow(points_df))
leafletProxy("map", data = points_df) %>% clearShapes() %>% addCircles()
})