Migration_flows / Data /__init__.py
ThGaskin's picture
Upload Data/__init__.py with huggingface_hub
ab70c85 verified
Raw
History Blame Contribute Delete
3.92 kB
import geopandas as gpd
import numpy as np
import os
import pandas as pd
import xarray as xr
from Plots.colors import colors
# We make relevant datasets contained in this folder available across the project, for easier integration and plotting.
# The base directory points to the folder containing this file:
base_dir = os.path.dirname(__file__)
# Make the ISO code lookup table available across the project. Also group countries by region.
lookup_table = pd.read_csv(os.path.join(base_dir, "Iso_code_lookup.csv"))
countries_by_region = pd.read_csv(os.path.join(base_dir, "countries_by_region.csv"))
countries_by_region = countries_by_region.groupby('Region').agg(list).to_dict()['Country ISO']
countries_by_region['Eastern Asia'].append('TWN')
countries_by_region['Middle East'] = countries_by_region.pop('Western Asia')
countries_by_region['South Asia'] = countries_by_region.pop('Southern Asia')
# Make the world shapefile available for plotting
world = gpd.read_file(os.path.join(base_dir, "world_shapefile/ne_50m_admin_0_countries.shp"))
world.to_crs("EPSG:3857", inplace=True) # Mercator projection by default
# Make the coordinates available. These are regular latitude/longitude coordinates
coordinates = pd.read_csv(os.path.join(base_dir, "Coordinates.csv"))[['Alpha-3 code', 'Latitude', 'Longitude']].set_index('Alpha-3 code').to_dict()
coordinates = dict((k, (coordinates['Latitude'][k], coordinates['Longitude'][k])) for k in coordinates['Latitude'].keys())
# Load the various comparison datasets into a dictionary for easy plotting, assigning consistent colours
# and markers for each.
flow_dsets = {
"Quantmig": dict(
data=xr.open_dataset(os.path.join(base_dir, "Flow_data/QuantMig_data/Quantmig_flows.nc")),
central='flow_50%', lower='flow_2.5%', upper='flow_97.5%',
primary_color=colors['c_orange'], secondary_color=colors['c_yellow']
),
"Statistics Sweden": dict(
data=xr.open_dataarray(os.path.join(base_dir, "Flow_data/National_Statistics/SWE_flows.nc")),
primary_color=colors['c_red'], secondary_color=colors['c_pink'], marker='v', s=10,
),
"Statistics Finland": dict(
data=xr.open_dataarray(os.path.join(base_dir, "Flow_data/National_Statistics/FIN_flows.nc")),
primary_color=colors['c_lightblue'], secondary_color=colors['c_darkblue'], marker='s', s=10
),
"StatNZ": dict(
data=xr.open_dataarray(os.path.join(base_dir, "Flow_data/National_Statistics/NZL_flows.nc")),
primary_color=colors['c_darkgreen'], secondary_color=colors['c_lightgreen'], marker='^', s=10
),
"Facebook data": dict(
data=xr.open_dataarray(os.path.join(base_dir, "Flow_data/Facebook/facebook_flows.nc")).where(lambda x: x>=25),
primary_color=colors['c_purple'], secondary_color='#E1C6EC', ec=colors['c_darkgrey'], marker='o', s=10, lw=0.5
)
}
# UN WPP data
population = 1e3 * xr.open_dataset(os.path.join(base_dir, "UN_WPP_data/UN_WPP_data.nc"))['Total Population, as of 1 January (thousands)']
death_rate = 1e-3 * xr.open_dataset(os.path.join(base_dir, "UN_WPP_data/UN_WPP_data.nc"))['Crude Death Rate (deaths per 1,000 population)']
WPP_net_migration = 1e3 * xr.open_dataset(os.path.join(base_dir, "UN_WPP_data/UN_WPP_data.nc"))['Net Number of Migrants (thousands)']
NatStat_net_migration = xr.open_dataarray(os.path.join(base_dir, "Net_migration/National_statistics.nc"))
WPP_data = xr.open_dataset(os.path.join(base_dir, "UN_WPP_data/UN_WPP_data.nc"))
# Calculate the fraction of population alive at start of 1990 in each destination country still alive at start of year
gamma = (1-death_rate.sel({"Year": range(1989, 2024)})).assign_coords({"Year": np.arange(1990, 2025, 1)}).rename({
"Country ISO": "Destination ISO"})
gamma.loc[{"Year": 1990}] = 1.0
gamma = gamma.cumprod('Year')
# Stock data
stock_data = xr.load_dataset(os.path.join(base_dir, "UN_stock_data/stock_data.nc"))