Migration map

data.py

import csv
import re
from pygeocoder import Geocoder

infile = open('mig_data.csv')
outfile = open('geo_data.csv','wb')
csvfile = list(csv.DictReader(infile))

fieldnames = ['origin_long', 'dest_long', 'dest_lat', 'MIGMOVE', 
              'Ref_Date', 'Coordinate','tempdest', 'Vector', 'Value', 
              'GEODEST','tempgeo', 'GEO', 'origin_lat']
csvwriter = csv.DictWriter(outfile, fieldnames=fieldnames)

unique = {}

for line in csvfile:
    # Non-robust way to get around Unicode import errors from French characters.
    line['GEO'] = line['GEO'].replace('\xe9','e')
    line['GEO'] = line['GEO'].replace('\xe8','e')
    line['GEODEST'] = line['GEODEST'].replace('\xe9','e')
    line['GEODEST'] = line['GEODEST'].replace('\xe8','e')
    
    # Pasting " City, Canada" makes sure we get the right coordinates
    line['tempgeo'] = re.split('[,\-/]',line['GEO'])[0] + " City, Canada"
    line['tempdest'] = re.split('[,\-/]',line['GEODEST'])[0] + " City, Canada"

    # After the previous steps we end up with "Non City, Canada" records (from
    # Non-Census Metropolitan Areas
    if (line['tempgeo'] not in unique.keys() or line['tempdest'] not in unique.keys()) \
    and line['tempgeo'] != "Non City, Canada" and line['tempdest'] != "Non City, Canada":
        unique[line['tempgeo']] = Geocoder.geocode(line['tempgeo']).coordinates
        unique[line['tempdest']] = Geocoder.geocode(line['tempdest']).coordinates
    
csvwriter.writeheader()
for line in csvfile:
    geo = line['tempgeo']
    geodest = line['tempdest']
    
    if geo != "Non City, Canada" and geodest != "Non City, Canada":
    
        line['origin_lat'] = unique[geo][0]
        line['origin_long'] = unique[geo][1]
        line['dest_lat'] = unique[geodest][0]
        line['dest_long'] = unique[geodest][1]
        
        csvwriter.writerow(line)
        
infile.close()
outfile.close()

migration_map.r

library(mapdata)
library(geosphere)
library(stringr)

setwd("~")

geodata = read.csv("geo_data.csv")

geodata$Value = as.numeric(as.character(geodata$Value))

geodata = cbind(as.data.frame(str_split_fixed(geodata$GEO,",",2)),geodata)
geodata = cbind(as.data.frame(str_split_fixed(geodata$GEODEST,",",2)),geodata)
names(geodata)[1] = "dest_city"
names(geodata)[2] = "dest_province"
names(geodata)[3] = "origin_city"
names(geodata)[4] = "origin_province"

keepcols = c("origin_city","dest_city","MIGMOVE","Value",
            "origin_lat","origin_long","dest_lat","dest_long", "Ref_Date")

geodata = geodata[keepcols]

majcmas = c("Vancouver","Calgary","Edmonton","Toronto","Ottawa-Gatineau","Montreal")

par(mfrow=c(2,3),mar=c(1,1,1,1),oma=c(0,0,0,0))
for (c in majcmas){
  map('worldHires','Canada',col="lightgrey",border="white", fill=TRUE,ylim=c(35,70))
  gsub = geodata[geodata$dest_city == c & geodata$MIGMOVE == "Net-migration" 
                 & geodata$Ref_Date == "2011",]
  dest_long = gsub$dest_long[1]
  dest_lat = gsub$dest_lat[1]

  for (i in 1:length(gsub$origin_city)){
    city1 = gsub[i,]
    gc = gcIntermediate(c(city1$origin_long,city1$origin_lat),c(dest_long,dest_lat),n=50,addStartEnd=TRUE)
    if (gsub[i,]$Value < 0) lines(gc,col="#0078BF",lwd=0.5)
    else lines(gc,col="red",lwd=0.5)
  }}