R ggvis date slider by month for interactive graph

Question

I'm attempting to make an interactive time series chart with ggvis, with a slider that will default using all data, but when sliding to right will cut the oldest data by month. Data is monthly, but I can not figure out how to make it work. Consider:

retail <- data.frame(Date = seq.Date(as.Date(parse_date_time("1/1/2007", "%m/%d/%y"))
                          , as.Date(parse_date_time("2/1/2017", "%m/%d/%y")),
                          by="month"),
                 Measure = rnorm(122))

u <- melt(retail, id="Date", measure = "Measure")

# Works for static graph
u %>%
  ggvis(~Date, ~value)

# Defaults to daily slider, no data displayed
u %>% 
  ggvis(~Date,~value) %>%
  layer_lines(input_slider(min(u$Date), max(u$Date), value=min(u$Date)))

I can not get the slider to control the Date variable, it keeps defaulting to daily. I noticed it defaults to daily in the documentation, but I can not find a workaround to force it to work with non-daily data.

Answer 1

For readability, let's create dataSlider outside:

library(lubridate) # for ceiling_date()

dateSlider <- input_slider(min=min(u$Date), max=max(u$Date), c(min(u$Date),
 max(u$Date)), map = function(x) ceiling_date(x, "month") )

map will change your arbitrary selected day from the input slider to the first day of the month. This mapped x then defines plot range.

u %>% 
 ggvis(~Date,~value) %>% 
  layer_lines() %>%
  scale_datetime("x", nice="month", domain = dateSlider, clamp = TRUE)

Edit: Removing white spaces before and after:

dateSlider2 <- input_slider(min=min(u$Date), max=max(u$Date), c(min(u$Date),
 max(u$Date)), map = function(x) floor_date(x, "month") )

u %>% 
 ggvis(~Date,~value) %>% 
  layer_lines() %>%
  scale_datetime("x", nice="month", expand=0,  domain = dateSlider2, 
   clamp = TRUE, override=TRUE)