This pull request is to address the following issues in the Highcharter function hctreemap:

New Dependencies

Both of these are included in the latest release of the tidyverse to CRAN. I got a R CMD Check Note for using NSE, which I resolved by adding

if (getRversion() >= "2.15.1") utils::globalVariables(c(".", "colorValue", "level", "name", "parent"))

to the highcharter-package.R file. This is new to me, so I figured if you have a preference for how to sort this out you’ll know how to change this better than I do.

Interface Change and Old Interface Deprecation

This pull request includes the function hctreemap2 which has the following interface

hctreemap2(data, group_vars, size_var, color_var = NULL, ...)

Note that this function accepts a dataframe rather than a treemap::treemap object. hctreemap remains unaltered, albeit with a deprecation warning.

The old interface looks like this

treemap(GNI2014,
        index = c("continent", "iso3"),
        vSize = "population",
        vColor = "GNI",
        type = "value",
        draw = FALSE) %>% 
  hctreemap(allowDrillToNode = TRUE, layoutAlgorithm = "squarified") %>% 
  hc_tooltip(pointFormat = "<b>{point.name}</b>:<br>
                            Pop: {point.value:,.0f}<br>
                            GNI: {point.valuecolor:,.0f}")
## Warning: 'hctreemap' is deprecated.
## Use 'hctreemap2' instead.
## See help("Deprecated")

and the new interface looks very similar

hctreemap2(data = GNI2014,
           group_vars = c("continent", "iso3"),
           size_var = "population",
           color_var = "GNI",
           layoutAlgorithm = "squarified",
           levelIsConstant = FALSE,
           levels = list(
             list(level = 1, dataLabels = list(enabled = TRUE)),
             list(level = 2, dataLabels = list(enabled = FALSE)),
             list(level = 3, dataLabels = list(enabled = FALSE))
           )) %>% 
  hc_colorAxis(minColor = brewer.pal(7, "Greens")[1],
               maxColor = brewer.pal(7, "Greens")[7]) %>% 
  hc_tooltip(pointFormat = "<b>{point.name}</b>:<br>
                            Pop: {point.value:,.0f}<br>
                            GNI: {point.colorValue:,.0f}")

Likely the biggest difference for users will be that they’ll need to specify the color palette themselves via hc_colorAxis rather than being able to make use of the palette argument in treemap::treemap. By default, hctreemap2 enables a color axis with minimum and maximum values determined by at the lowest depth in the treemap, in keeping with treemap::treemap. I’ve included the above as an example in the function documentation, so hopefully people will get the hint about how to use RColorBrewer to make their lives better.

If the user doesn’t specify a variable for the color mapping, the variable for the size mapping is used by default. Additionally, allowDrillToNode defaults to TRUE.

We can make a rather boring three level treemap to demonstrate that three level treemaps work

data.frame(index1 = sample(LETTERS[1:3], 500, replace = T),
           index2 = sample(LETTERS[4:6], 500, replace = T),
           index3 = sample(LETTERS[7:9], 500, replace = T),
           value = rpois(500, 5)) %>%
  hctreemap2(
    group_vars = c("index1", "index2", "index3"),
    size_var = "value"
  )

Non-Unique Cells Names Throw An Error

set.seed(27)

data.frame(
  index1 = sample(LETTERS[1:6], 500, replace = T),
  index2 = sample(LETTERS[1:6], 500, replace = T),
  value = abs(rnorm(500))) %>% 
  hctreemap2(c("index1", "index2"), "value")
## Error in hctreemap2(., c("index1", "index2"), "value"): Treemap data uses same label at multiple levels.

The error checking also catches missing data:

data.frame(
  index1 = c(1, NA),
  index2 = c(NA, 2),
  value = abs(rnorm(500))) %>% 
  hctreemap2(c("index1", "index2"), "value")
## Error in hctreemap2(., c("index1", "index2"), "value"): Treemap data uses same label at multiple levels.

This forces users to recode their data before continuing and should allow arbitrary depth treemaps. Lastly, a single level treemap just as another example

hctreemap2(data = GNI2014,
           group_vars = "continent",
           size_var = "population") %>% 
  hc_colorAxis(minColor = brewer.pal(7, "Greens")[1],
               maxColor = brewer.pal(7, "Greens")[7]) %>% 
  hc_tooltip(pointFormat = "<b>{point.name}</b>:<br>
                            Pop: {point.value:,.0f}<br>
                            GNI: {point.colorValue:,.0f}")