This commit is contained in:
2025-09-26 00:11:14 -05:00
commit 3f6cd4ba59
14 changed files with 688056 additions and 0 deletions

127
mappings.py Normal file
View File

@@ -0,0 +1,127 @@
# This dictionary will be used to ensure standardized country naming conventions for each dataset.
country_name_map = {
'Puerto Rico' : 'United States',
'Guam' : 'United States',
'Netherlands Antilles' : 'Netherlands',
'Jersey' : 'United Kingdom',
'Greenland' : 'Netherlands',
'New Caledonia' : 'France',
'Guernsey' : 'United Kingdom',
'Congo (Kinshasa)' : 'Democratic Republic of the Congo',
'Congo (Brazzaville)' : 'Republic of the Congo',
'Reunion' : 'France',
"Martinique" : "France",
"Guadeloupe" : "France",
"French Guiana" : "France",
"Gibraltar" : "United Kingdom",
"French Polynesia" : "France",
"Isle of Man" : "United Kingdom",
"Bermuda" : "United Kingdom",
'Anguilla': 'United Kingdom',
'Saint Pierre and Miquelon': 'France',
'Wallis and Futuna': 'France',
'Aruba': 'Netherlands',
'Faroe Islands': 'Denmark',
'Cook Islands': 'New Zealand',
'Mayotte': 'France',
"Northern Mariana Islands" : "United States",
"Réunion" : "France",
'United States of America': 'United States',
'Türkiye' : 'Turkey',
'Czechia' : 'Czech Republic',
'Burma/Myanmar' : 'Burma',
'Ivory Coast' : "Cote d'Ivoire",
"Timor-Leste" : "East Timor",
"North Macedonia" : "Macedonia",
'Turkiye': 'Turkey',
'Russian Federation' : 'Russia',
"Taiwan, China" : "Taiwan",
"Venezuela, Bolivarian Rep. of" : "Venezuela",
'Korea, Rep.' : 'South Korea',
"Korea, Dem. People's Rep." : 'North Korea',
"Hong Kong SAR, China" : "Hong Kong",
"Macao SAR, China" : "Macau",
"Egypt, Arab Rep.": "Egypt",
"Viet Nam" : "Vietnam",
"South Sudan, The Republic of" : "South Sudan",
"Sudan, The Republic of" : "Sudan",
"Korea, Republic of" : "South Korea",
"Macedonia, The former Yugoslav Rep. of" : "Macedonia",
"Venezuela, RB" : "Venezuela",
"Bahamas, The" : "Bahamas",
"Iran, Islamic Rep." : "Iran",
"Congo, Democratic Republic of the" : "Democratic Republic of the Congo",
"Myanmar" : "Burma",
"Kyrgyz Republic" : "Kyrgyzstan",
"Yemen, Rep." : "Yemen",
"Congo, Dem. Rep." : "Democratic Republic of the Congo",
"Tanzania, United Republic of" : "Tanzania",
"Iran, Islamic Rep. of" : "Iran",
"Congo, Rep." : "Republic of the Congo",
"Lao PDR" : "Laos",
"Slovak Republic" :"Slovakia",
"Gambia, The" : "Gambia",
'Western Sahara': 'Morocco',
'Venezuela (Bolivarian Republic of)': 'Venezuela',
'Iran (Islamic Republic of)': 'Iran',
'Congo': 'Republic of the Congo',
'Congo (Democratic Republic of the)': 'Democratic Republic of the Congo',
'Lao People\'s Democratic Republic': 'Laos',
'Republic of Moldova': 'Moldova',
'Syrian Arab Republic': 'Syria',
'Tanzania (United Republic of)': 'Tanzania',
"Hong Kong, China (SAR)" : "Hong Kong",
"Hong Kong, China" : "Hong Kong",
"Korea (Democratic People's Rep. of)" : "North Korea",
"Korea (Republic of)" : "South Korea",
"Côte d'Ivoire" : "Cote d'Ivoire",
"Bolivia (Plurinational State of)" : "Bolivia",
"Moldova (Republic of)" : "Moldova",
"Brunei Darussalam" : "Brunei",
"Cabo Verde" : "Cape Verde",
'Turkiye': 'Turkey', # Added both versions to be safe
'Russian Federation' : 'Russia',
'Korea, Rep.' : 'South Korea',
"Korea, Dem. People's Rep." : 'North Korea',
"Hong Kong SAR, China" : "Hong Kong",
"Macao SAR, China" : "Macau",
"Czechia" : "Czech Republic",
"Egypt, Arab Rep.": "Egypt",
"Viet Nam" : "Vietnam",
"Venezuela, RB" : "Venezuela",
"Bahamas, The" : "Bahamas",
"Iran, Islamic Rep." : "Iran",
"Myanmar" : "Burma",
"Kyrgyz Republic" : "Kyrgyzstan",
"Yemen, Rep." : "Yemen",
"Congo, Dem. Rep." : "Democratic Republic of the Congo",
"Congo, Rep." : "Republic of the Congo",
"Lao PDR" : "Laos",
"North Macedonia" : "Macedonia",
"Slovak Republic" :"Slovakia",
"Gambia, The" : "Gambia",
'Western Sahara': 'Morocco',
'Republic of Korea': 'South Korea',
'Czechia': 'Czech Republic',
'Viet Nam': 'Vietnam',
'Venezuela (Bolivarian Republic of)': 'Venezuela',
'Iran (Islamic Republic of)': 'Iran',
'Myanmar': 'Burma',
'Congo': 'Republic of the Congo',
'Congo (Democratic Republic of the)': 'Democratic Republic of the Congo',
'Lao People\'s Democratic Republic': 'Laos',
'Republic of Moldova': 'Moldova',
'Syrian Arab Republic': 'Syria',
'Tanzania (United Republic of)': 'Tanzania',
"Hong Kong, China (SAR)" : "Hong Kong",
"Korea (Democratic People's Rep. of)" : "North Korea",
"Korea (Republic of)" : "South Korea",
"Türkiye" : "Turkey",
"Côte d'Ivoire" : "Cote d'Ivoire",
"Bolivia (Plurinational State of)" : "Bolivia",
"North Macedonia" : "Macedonia",
"Moldova (Republic of)" : "Moldova",
"Brunei Darussalam" : "Brunei",
"Timor-Leste" : "East Timor",
"Cabo Verde" : "Cape Verde"
}