Add shortcut for density plots

[nuno-agostinho]

Hey there @jbkunst!

I created the following function to get density plots from density objects:

hchart.density <- function (..., area = FALSE) {
    type <- ifelse(area, "area", "spline")
    args <- list(...)
    data <- lapply(args, function(i) {
        x <- i$x
        y <- i$y
        list.parse3(data.frame(cbind(x = x, y = y)))
    })
    hc <- highchart()
    for (d in data)
        hc <- hc %>% hc_add_series(data = d, type = type)
    return(hc)
}

Which would make it simple to plot multiple density objects.

d1 <- density(mtcars$mpg[1:32/2])
d2 <- density(mtcars$mpg[32/2:32])
d3 <- density(mtcars$mpg[5:(16+5)])
hchart(d1, d2, d3, area = TRUE)

However, I thought the function does not share the same philosophy you apply to the other hchart methods. In that sense, I tried to make it more like one of your shortcuts:

hc_add_series_density <- function (hc, x, area = FALSE, ...) {
    type <- ifelse(area, "area", "spline")
    data <- list.parse3(data.frame(cbind(x = x$x, y = x$y)))
    return(hc %>% hc_add_series(data = data, type = type))
}

And the use would be more like this:

highchart() %>%
    hc_add_series_density(d1, area = TRUE) %>%
    hc_add_series_density(d2, area = TRUE) %>%
    hc_add_series_density(d3, area = TRUE) 

But after coming with this, I thought this was maybe too simple to add to your package. So what do you think? Should I add the code of hc_add_series_density in a pull request?

[jbkunst]

Hi @nuno-agostinho ,

Really nice addition. It's look very nice.

I think we can do:

• The helper: hc_add_series_density here. Can we add in this function the lineif(is.numeric(x)) x <- density(x)?. I think it would be nice give a numeric object which I want to show as a density.
• Then the generic function hchart.density <- function (object,..., area = FALSE) { hc_add_series_density(highcharts(), object, ...)} here
• And finally the shortcut hcdensity <- function(x, ...) { hchart.density(x, ...)} here due highcharts version for common plots #99

So we can do:

x <- rnorm(1000)
y <- rexp(1000)

hcdensity(x) %>%
  hc_add_series_density(y)

# and
hcdensity(density(x)) %>% # not the correct usage but its possible.
  hc_add_series_density(density(y))

# or
hchart(density(x)) %>%
  hc_add_series_density(density(y))

What do you think?

Thanks :D

[nuno-agostinho]

I think your tips are great! 👍 I've included them in the pull request #109.

I'll close this issue now.