I want to improve the appearance of my highcharter sankey diagram.
The labels of the nodes should be left and right to the plot and not "on" the nodes. So in the plot shown below the country names should appear left and right to the plot.
Additionally I am failing to add a thousand separator "," to the labels showing up when hovering over links and nodes.
This is the code I use for the sankey plot:
highchart() %>%
hc_add_series(
type = "sankey",
data = df_list_spaces,
nodes = nodes_list,
links = df_list_spaces,
name = "assets",
dataLabels = list(
enabled = TRUE
))
And this is how it looks:
To add a thousand separator I tried this:
# change thousands sep
hcoptslang <- getOption("highcharter.lang")
hcoptslang$thousandsSep <- ","
options(highcharter.lang = hcoptslang)
This works for other highcharter objects but strangely not for the "sankey" chart.
To reproduce the plot I also add the code with all the data transformation needed. There are probably nicer ways to end up with the final data for the plot but this is how I did it:
library(tidyverse)
library(highcharter)
df <- tibble(from = c("Brazil", "Brazil", "Poland", "Spain", "Poland"),
toto = c("Portugal", "Spain", "Portugal", "Brazil", "Poland"),
assets = c(5324523.4523, 232452345.234, 2345234.52345, 3234523.45234, 63452345.2345))
# bring rows into lists
# add space at the end for values of "to"
transform_to_list <- function(data, from, to, weight, add_spaces = FALSE) {
df_list <- list()
for (i in 1:nrow(data)) {
df_list[[i]] <- list(from = data[[from]][i],
# add spaces, so there is a difference between "from" and "to" countries
to = ifelse(test = add_spaces == TRUE,
yes = paste0(data[[to]][i], " "),
no = data[[to]][i]
),
weight = data[[weight]][i]
)
}
return(df_list)
}
# select the data set, from, to and weight column; add_spaces = TRUE is needed for Sankey diagramms,
# for other things, for example "dependency_wheels" add_spaces should be FALSE
df_list_spaces <- transform_to_list(data = df, from = "from", "toto", weight = "assets", add_spaces = TRUE)
#--------------------------------------
# next step is to attach colors to the countries, but the problem is that we had to add spaces to the "to"
# countries and hence we have to find a way to attach the same color for example for "Brazil" and "Brazil ".
# We do this with a little function:
# inputs: data, from column, to column and a colorpallete
transform_to_nodes_list <- function(data, from, to, colors) {
# create a vector of unique values from the "from" column
from_col <- unique(data[[from]])
# create a vector of values from the "to" column, with a space added at the end
to_col <- unique(paste0(data[[to]], " "))
# combine the "from" and "to" columns into a single vector
id_col <- c(from_col, to_col)
# create a "helper" column that ignores spaces at the end of values in the "id" column
helper_col <- gsub("\\s+$", "", id_col)
# match colors based on helper column, first create factors,
# then change class to numeric so first factor gets first color of the color palette, second factor second color..
color_col <- colors[as.numeric(factor(helper_col))]
# not necessary step
d0 <- tibble(id_col, color_col)
nodes_list <- list()
for (i in 1:nrow(d0)) {
# from <- deparse(substitute(from))
# to <- deparse(substitute(to))
# weight <- deparse(substitute(weight))
nodes_list[[i]] <- list(id = d0[["id_col"]][i],
color = d0[["color_col"]][i])
}
return(nodes_list)
}
# color palette:
pal <- c("brown", "blue", "yellow", "green", "orange", "red")
# call the functions:
df_list_spaces <- transform_to_list(data = df, from = "from", "toto", weight = "assets", add_spaces = TRUE)
nodes_list <- transform_to_nodes_list(data = df, from = "from", to = "toto", colors = pal)
# and finally plot
highchart() %>%
hc_add_series(
type = "sankey",
data = df_list_spaces,
nodes = nodes_list,
links = df_list_spaces,
name = "assets",
dataLabels = list(
enabled = TRUE
))
