Example of using function to create d3 sankey diagram from multiple categorical variables + one numerical value variable of tidy data frame

sankey.png

sankey.R

library(tidyverse)
library(tidygraph)
library(igraph)
library(networkD3)

# https://www.kaggle.com/unitednations/refugee-data/data
asylum_seekers_raw <- read_csv("asylum_seekers.csv")

# get the top 10 countries of origin by 'Total decisions'
top_orig <- asylum_seekers_raw %>%
  filter(Year == 2016) %>%  
  group_by(Origin) %>% 
  summarise(total = sum(`Total decisions`, na.rm = TRUE)) %>% 
  top_n(10, total) %>% 
  dplyr::pull(1)
  
# list of destination countries we want to look at
euro_countries <- 
  c("Germany", "Sweden", "France", "United Kingdom", "Netherlands", "Switzerland", "Italy", "Belgium")

# filter and clean the data we want to visualise
asylum_seekers <- asylum_seekers_raw %>% 
  filter(Origin %in% top_orig, `Country / territory of asylum/residence` %in% euro_countries) %>%
  select(1, 3, 2, 8, 10) %>% 
  filter(Year == 2016) %>% 
  select(-Year) %>% 
  drop_na() %>% 
  gather(Decision, Weight, 3:4)

############################################## SANKEY FUNCTION ######################################
sankey_func <- function(data, val_col, ...) {
  
  weight <- enquo(val_col)
  
  df <- 1:(ncol(data)-2) %>% 
    # collapse dataframe into 3 columns: from, to, weight
    map_df(~ select(data, from = !! quo(names(data)[.x]), to = !! quo(names(data)[.x + 1]), !! weight)) %>% 
    drop_na() %>% 
    group_by(from, to) %>% 
    summarise(weight = sum(!! weight)) %>% 
    mutate(colour = to)
  
  ig <- igraph::graph_from_data_frame(df) %>% 
    as_tbl_graph()
  
  nodes <- as_tibble(ig) %>% 
    rowid_to_column("id") %>% 
    mutate(id = id -1) %>% 
    as.data.frame
  
  edges <- ig %>% 
    activate(edges) %>% 
    as_tibble() %>% 
    mutate(from = from - 1, to = to - 1) %>% 
    as.data.frame
  
  sankeyNetwork(Links = edges, Nodes = nodes, Source = "from", Target = "to", 
                NodeID = "name", Value = "weight", LinkGroup = "colour", 
                fontSize = 14, fontFamily = "Roboto")
}
########################################## END OF FUNCTION ##########################################

# run the function on the cleaned asylum seeker data
sankey_func(data = asylum_seekers, val_col = Weight)