Visualising The Evolution Of Migration Flows With rCharts

Heaven we hope is just up the road (Atlas, Coldplay)

Following with the analysis of migration flows, I have done next two visualizations. These charts are called bump charts and are very suitable to represent rankings. This is what I have done:

  • Obtaining top 20 countries of the world according to % of migrants respect its population
  • To do this, I divide total number of migrants between 1960 and 2009 by the mean population in the same period.
  • I do the same to obtain top 20 countries of the world according to % of immigrants.
  • In both cases, I only consider countries with population greater than 2 million.
  • For these countries, I calculate % of migrants in each decade (60’s, 70’s, 80’s, 90’s and 00’s), dividing total number of migrants by mean population each decade
  • I do the same in the case of immigrants.
  • Instead of representing directly % of migrants and immigrants, I represent the ranking of countries according these indicators by decade

This is the bump chart of migrants:

And this is the one of immigrants:

Some comments:

  • There is a permanent exodus in Puerto Rico: all decades (except 70’s) is located in the top 1 of countries with most migrants respect its population
  • Ireland is also living a diaspora although in the 00’s decade has lost some positions
  • Albania, Georgia and Bosnia and Herzegovina are gaining positions. Is East Europe gradually becoming uncomfortable?
  • Jamaica is also moving up in this sad competition.
  • On the other hand, Hong Kong and Israel are persistently leaders as receivers
  • Saudi Arabia has presented an impressive growth receiving immigrants since 70’s
  • United States does not appear in the immigrants ranking
  • Singapore is gaining positions: in the 00’s decade is the third receiver country
  • Also in the 00s, Switzerland is the first European country in the ranking, holding the fifth position

I like using rCharts as well as using Enigma data sets, as I have done previously. This is the code:

library(data.table)
library(rCharts)
library(dplyr)
setwd("YOUR WORKING DIRECTORY HERE")
populflows = read.csv(file="enigma-org.worldbank.migration-remittances.migrants.migration-flow-c57405e33412118c8757b1052e8a1490.csv", stringsAsFactors=FALSE)
population = fread("enigma-org.worldbank.hnp.data-eaa31d1a34fadb52da9d809cc3bec954.csv")
population %>% 
  filter(indicator_name=="Population, total") %>% 
  as.data.frame %>% 
  mutate(decade=(year-year%%10)) %>% 
  group_by(country_name, country_code, decade) %>% 
  summarise(population=mean(value)) %>% 
  plyr::rename(., c("country_name"="country")) -> population2
populflows %>% filter(!is.na(total_migrants)) %>% 
  group_by(migration_year, destination_country) %>% 
  summarise(inmigrants = sum(total_migrants))  %>% 
  plyr::rename(., c("destination_country"="country", "migration_year"="decade"))   -> inmigrants
populflows %>% filter(!is.na(total_migrants)) %>% 
  group_by(migration_year, country_of_origin) %>% 
  summarise(migrants = sum(total_migrants)) %>%  
  plyr::rename(., c("country_of_origin"="country", "migration_year"="decade"))   -> migrants
# Join of data sets
migrants %>% 
  merge(inmigrants, by = c("country", "decade")) %>%
  merge(population2, by = c("country", "decade")) %>%
  mutate(p_migrants=migrants/population, p_inmigrants=inmigrants/population) -> populflows2
# Global Indicators
populflows2 %>% 
  group_by(country) %>% 
  summarise(migrants=sum(migrants), inmigrants=sum(inmigrants), population=mean(population)) %>% 
  mutate(p_migrants=migrants/population, p_inmigrants=inmigrants/population)  %>% 
  filter(population > 2000000)  %>%
  mutate(rank_migrants = dense_rank(desc(p_migrants)), rank_inmigrants = dense_rank(desc(p_inmigrants))) -> global
# Migrants dataset
global %>% 
  filter(rank_migrants<=20) %>% 
  select(country) %>% 
  merge(populflows2, by = "country") %>% 
  arrange(decade, p_migrants) %>%
  mutate(decade2=as.numeric(as.POSIXct(paste0(as.character(decade), "-01-01"), origin="1900-01-01"))) %>%
  plyr::ddply("decade", transform, rank = dense_rank(p_migrants)) -> migrants_rank
# Migrants dataset
global %>% 
  filter(rank_inmigrants<=20) %>% 
  select(country) %>% 
  merge(populflows2, by = "country") %>% 
  arrange(decade, p_inmigrants) %>%
  mutate(decade2=as.numeric(as.POSIXct(paste0(as.character(decade), "-01-01"), origin="1900-01-01"))) %>%
  plyr::ddply("decade", transform, rank = dense_rank(p_inmigrants)) -> inmigrants_rank
# Function for plotting
plotBumpChart <- function(df){
  bump_chart = Rickshaw$new()
  mycolors = ggthemes::tableau_color_pal("tableau20")(20)
  bump_chart$layer(rank ~ decade2, group = 'country_code', data = df, type = 'line', interpolation = 'none', colors = mycolors)
  bump_chart$set(slider = TRUE, highlight = TRUE, legend=TRUE)
  bump_chart$yAxis(tickFormat = "#!  function(y) { if (y == 0) { return '' } else { return String((21-y)) } } !#")
  bump_chart$hoverDetail(yFormatter = "#! function(y){return (21-y)} !#")
  return(bump_chart)
}
plotBumpChart(migrants_rank)
plotBumpChart(inmigrants_rank)

4 thoughts on “Visualising The Evolution Of Migration Flows With rCharts

    1. This is because these files are stored locally in my working directory. Register in enigma.io and download them. Once you have done this you can point them using my code. (take into account that names will be different of mine)

  1. then, it is not a good example to replicate.
    I will look for another rChart post with replicable codes.

    thanks very much for your reply.

Leave a Reply to aschinchon Cancel reply

Your email address will not be published. Required fields are marked *