.Rhistory

View(trade.dist.matrix)
trade.dist.matrix <- trade.dist.matrix[, -1]
View(trade.dist.matrix)
trade.dist.matrix <- t(trade.dist.matrix)
View(trade.dist.matrix)
trade.dist.matrix$country_codes <- rownames(trade.dist.matrix)
View(ruhlen2)
# Get geograhic data for each country (longitude and latitude)
geo <- read.csv("./input_data/geo.csv")
# Get 2008 levels of income for each country
mad08 <- read.csv("./input_data/maddison_subset_2008gdp.csv")
if (add.trade) {
# Trade distance matrix - direct input method (i.e. data is already cleaned up in Excel)
trade.dist.matrix <- read.csv("./input_data/trade_distance_matrix.csv")
colnames(trade.dist.matrix)[1] <- "country_code"
}
# Harmonize all datasets by selecting country rows that they all have
full <- merge(ruhlen2, mad, by = "country_code")
full <- merge(full, geo, by = "country_code")
if (add.trade) {
full <- merge(full, trade.dist.matrix, by = "country_code") ############ Trade
}
full <- select(full, country_code)
ruhlen2 <- merge(full, ruhlen2, by = 'country_code')
mad2 <- merge(full, mad, by = 'country_code')
geo2 <- merge(full, geo, by = 'country_code')
mad08 <- merge(full, mad08, by = 'country_code')
if (add.trade) {
trade.dist.matrix <- merge(full, trade.dist.matrix, by = 'country_code') ############ Trade
}
identical(as.vector(mad2$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
identical(as.vector(geo2$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
if (add.trade) {
identical(as.vector(trade.dist.matrix$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
}
rownames(trade.dist.matrix) <- trade.dist.matrix[, 1]
trade.dist.matrix <- trade.dist.matrix[, -1]
trade.dist.matrix <- as.data.frame(t(trade.dist.matrix))
View(trade.dist.matrix)
trade.dist.matrix$country_codes <- rownames(trade.dist.matrix)
View(trade.dist.matrix)
trade.dist.matrix <- merge(full, trade.dist.matrix, by = 'country_code')
trade.dist.matrix$country_codes
# Get geograhic data for each country (longitude and latitude)
geo <- read.csv("./input_data/geo.csv")
# Get 2008 levels of income for each country
mad08 <- read.csv("./input_data/maddison_subset_2008gdp.csv")
if (add.trade) {
# Trade distance matrix - direct input method (i.e. data is already cleaned up in Excel)
trade.dist.matrix <- read.csv("./input_data/trade_distance_matrix.csv")
colnames(trade.dist.matrix)[1] <- "country_code"
}
# Harmonize all datasets by selecting country rows that they all have
full <- merge(ruhlen2, mad, by = "country_code")
full <- merge(full, geo, by = "country_code")
if (add.trade) {
full <- merge(full, trade.dist.matrix, by = "country_code") ############ Trade
}
full <- select(full, country_code)
ruhlen2 <- merge(full, ruhlen2, by = 'country_code')
mad2 <- merge(full, mad, by = 'country_code')
geo2 <- merge(full, geo, by = 'country_code')
mad08 <- merge(full, mad08, by = 'country_code')
if (add.trade) {
trade.dist.matrix <- merge(full, trade.dist.matrix, by = 'country_code') ############ Trade
}
identical(as.vector(mad2$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
identical(as.vector(geo2$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
if (add.trade) {
identical(as.vector(trade.dist.matrix$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
}
rownames(trade.dist.matrix) <- trade.dist.matrix[, 1]
trade.dist.matrix <- trade.dist.matrix[, -1]
trade.dist.matrix <- as.data.frame(t(trade.dist.matrix))
trade.dist.matrix$country_code <- rownames(trade.dist.matrix)
trade.dist.matrix <- merge(full, trade.dist.matrix, by = 'country_code')
View(trade.dist.matrix)
rownames(trade.dist.matrix) <- trade.dist.matrix[, 1]
trade.dist.matrix <- trade.dist.matrix[, -1]
View(trade.dist.matrix)
isSymmetric(trade.dist.matrix)
isSymmetric(as.matrix(trade.dist.matrix))
class(trade.dist.matrix[5, 6])
class(trade.dist.matrix[5, 60])
mode(trade.dist.matrix) <- "numeric"
diag(trade.dist.matrix)
trade.dist.matrix[5, 5]
class(trade.dist.matrix[5, 5])
diag(trade.dist.matrix)
class(trade.dist.matrix)
View(trade.dist.matrix)
z <- as.dist(trade.dist.matrix)
trade.dist.matrix <- as.dist(trade.dist.matrix)
as.matrix(trade.dist.matrix)
diag(as.matrix(trade.dist.matrix))
View(as.matrix(trade.dist.matrix))
class(trade.dist.matrix)
ncf::mantel.test(as.matrix(lang.dist.matrix), as.matrix(trade.dist.matrix), resamp = 9999)
# Create growth distance matrix based on Maddison income data using Euclidean distance
mad3 <- mad2[, 3:ncol(mad2)] # Extract the growth numbers
rownames(mad3) <- mad2$country_code # Rename rows using country codes
mad3 <- t(scale(t(mad3))) # Standardize growth rates within each country so that mean = 0, SD = 1
growth.dist.matrix <- vegdist(mad3, "euclidean")
# Create language distance matrix for Ruhlen linguistic data using Jaccard distance
ruhlen3 <- ruhlen2[, 12:ncol(ruhlen2)] # Extract phoneme presence/absence for all 728 phonemes
rownames(ruhlen3) <- ruhlen2$country_code # Rename rows to ISO country codes
lang.dist.matrix <- vegdist(ruhlen3, "jaccard", binary = TRUE)
# Create geographic distance matrix for geographic data using great circle distance
geo3 <- geo2[ , 2:3]
geo.dist.matrix <- rdist.earth(geo3, geo3)
rownames(geo.dist.matrix) <- geo2[, 1]
colnames(geo.dist.matrix) <- geo2[, 1]
geo.dist.matrix <- log(as.dist(geo.dist.matrix)) # Log distance
# Create distance matrix for 2008 levels of income
mad08_2 <- as.matrix(mad08$X2008) # Extract the income numbers
rownames(mad08_2) <- mad08$country_code # Rename rows using country codes
mad08.dist.matrix <- log(vegdist(mad08_2, "euclidean")) # Log distance
ncf::mantel.test(as.matrix(lang.dist.matrix), as.matrix(trade.dist.matrix), resamp = 9999)
create.dist.matrices <- function (mad, add.trade=FALSE) {
# Get linguistic data from Ruhlen
ruhlen <- read.csv("./input_data/ruhlen.csv")
ruhlen <- rename(ruhlen, country_code = Country..ISO.Alpha.3.Code.) # Clean up column name for country code
# Get most spoken language for each country
ruhlen2 <- ruhlen %>%
group_by(country_code) %>%
arrange(desc(Population)) %>%
slice(1) %>% # In case there's more than one max language, just take the first one
ungroup
## Dealing with missing languages
# The real dominant languages of certain countries are missing
# E.g. For Australia, there's no Australian English and so the dominant language is (incorrectly) a small aboriginal language
# In this section, I manually substitute the phonemic data for some countries' languages with others that make sense
# E.g. Substitute Australia with the phonemic data from Britain's English
ruhlen2[which(ruhlen2$country_code == "AFG"), 12:739] <- ruhlen[which(ruhlen$ISO_A3 == "pbt-AFG"), 12:739] # Pashto (not Farsi) is the largest language in Afghanistan
ruhlen2[which(ruhlen2$country_code == "AUS"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "GBR"), 12:739] # Australia <- Britain
ruhlen2[which(ruhlen2$country_code == "ARG"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Argentina <- Colombia
ruhlen2[which(ruhlen2$country_code == "BOL"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Bolivia <- Colombia
ruhlen2[which(ruhlen2$country_code == "BWA"), 12:739] <- ruhlen[which(ruhlen$ISO_A3 == "tsn-ZAF"), 12:739] # Tswana is main language of Botswana (but categorized as a language of South Africa)
ruhlen2[which(ruhlen2$country_code == "CHE"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "DEU"), 12:739] # Switzerland <- Germany
ruhlen2[which(ruhlen2$country_code == "CHL"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Chile <- Colombia
ruhlen2[which(ruhlen2$country_code == "GTM"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "CRI"), 12:739] # Guatemala <- Costa Rica
ruhlen2[which(ruhlen2$country_code == "HND"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "CRI"), 12:739] # Honduras <- Costa Rica
ruhlen2[which(ruhlen2$country_code == "NIC"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "CRI"), 12:739] # Nicaragua <- Costa Rica
ruhlen2[which(ruhlen2$country_code == "NZL"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "GBR"), 12:739] # New Zealand <- Britain
ruhlen2[which(ruhlen2$country_code == "PER"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Peru <- Colombia
ruhlen2[which(ruhlen2$country_code == "PRY"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Paraguay <- Colombia
ruhlen2[which(ruhlen2$country_code == "SDN"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "EGY"), 12:739] # Sudan <- Egypt
ruhlen2[which(ruhlen2$country_code == "SLV"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "CRI"), 12:739] # El Salvador <- Costa Rica
ruhlen2[which(ruhlen2$country_code == "VEN"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Venezuela <- Colombia
# Get geograhic data for each country (longitude and latitude)
geo <- read.csv("./input_data/geo.csv")
# Get 2008 levels of income for each country
mad08 <- read.csv("./input_data/maddison_subset_2008gdp.csv")
if (add.trade) {
# Trade distance matrix - direct input method (i.e. data is already cleaned up in Excel)
trade.dist.matrix <- read.csv("./input_data/trade_distance_matrix.csv")
colnames(trade.dist.matrix)[1] <- "country_code"
}
# Harmonize all datasets by selecting country rows that they all have
full <- merge(ruhlen2, mad, by = "country_code")
full <- merge(full, geo, by = "country_code")
if (add.trade) {
full <- merge(full, trade.dist.matrix, by = "country_code") ############ Trade
}
full <- select(full, country_code)
ruhlen2 <- merge(full, ruhlen2, by = 'country_code')
mad2 <- merge(full, mad, by = 'country_code')
geo2 <- merge(full, geo, by = 'country_code')
mad08 <- merge(full, mad08, by = 'country_code')
if (add.trade) {
trade.dist.matrix <- merge(full, trade.dist.matrix, by = 'country_code') ############ Trade
}
identical(as.vector(mad2$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
identical(as.vector(geo2$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
if (add.trade) {
identical(as.vector(trade.dist.matrix$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
}
if (add.trade) {
rownames(trade.dist.matrix) <- trade.dist.matrix[, 1]
trade.dist.matrix <- trade.dist.matrix[, -1]
trade.dist.matrix <- as.data.frame(t(trade.dist.matrix))
trade.dist.matrix$country_code <- rownames(trade.dist.matrix)
trade.dist.matrix <- merge(full, trade.dist.matrix, by = 'country_code')
rownames(trade.dist.matrix) <- trade.dist.matrix[, 1]
trade.dist.matrix <- trade.dist.matrix[, -1]
trade.dist.matrix <- as.dist(trade.dist.matrix)
}
# Create growth distance matrix based on Maddison income data using Euclidean distance
mad3 <- mad2[, 3:ncol(mad2)] # Extract the growth numbers
rownames(mad3) <- mad2$country_code # Rename rows using country codes
mad3 <- t(scale(t(mad3))) # Standardize growth rates within each country so that mean = 0, SD = 1
growth.dist.matrix <- vegdist(mad3, "euclidean")
# Create language distance matrix for Ruhlen linguistic data using Jaccard distance
ruhlen3 <- ruhlen2[, 12:ncol(ruhlen2)] # Extract phoneme presence/absence for all 728 phonemes
rownames(ruhlen3) <- ruhlen2$country_code # Rename rows to ISO country codes
lang.dist.matrix <- vegdist(ruhlen3, "jaccard", binary = TRUE)
# Create geographic distance matrix for geographic data using great circle distance
geo3 <- geo2[ , 2:3]
geo.dist.matrix <- rdist.earth(geo3, geo3)
rownames(geo.dist.matrix) <- geo2[, 1]
colnames(geo.dist.matrix) <- geo2[, 1]
geo.dist.matrix <- log(as.dist(geo.dist.matrix)) # Log distance
# Create distance matrix for 2008 levels of income
mad08_2 <- as.matrix(mad08$X2008) # Extract the income numbers
rownames(mad08_2) <- mad08$country_code # Rename rows using country codes
mad08.dist.matrix <- log(vegdist(mad08_2, "euclidean")) # Log distance
output <- list(growth.dist.matrix, lang.dist.matrix, geo.dist.matrix, mad08.dist.matrix)
return(output)
}
matrices <- create.dist.matrices(mad_subset5, add.trade=TRUE)
mad_subset5 <- read.csv("./input_data/maddison_subset5.csv") # 1950-2008, 137 countries, 5-year CAGR ----- PRIMARY OPTION
# These are for checking only
mad <- read.csv("./input_data/maddison_subset4a.csv") # 1950-2000, 138 countries, 10-year CAGR
mad <- read.csv("./input_data/maddison_subset6.csv") # random (check on Mantel test)
mad <- read.csv("./input_data/maddison_subset7.csv") # first half of countries are randomized (check on Mantel test)
# By region
mad_subset_europe <- read.csv("./input_data/maddison_subset_europe.csv") # 1950-2008, 23 countries, 5-year CAGR, Europe only
mad_subset_africa <- read.csv("./input_data/maddison_subset_africa.csv") # 1950-2008, 48 countries, 5-year CAGR, Sub-Saharan Africa only
mad_subset_eastasia <- read.csv("./input_data/maddison_subset_eastasia.csv") # 1950-2008, 15 countries, 5-year CAGR, East Asia only
mad_subset_latam <- read.csv("./input_data/maddison_subset_latam.csv") # 1950-2008, 23 countries, 5-year CAGR, Latin America only
mad_subset_anglo <- read.csv("./input_data/maddison_subset_anglo.csv") # 1950-2008, 6 countries, 5-year CAGR, Anglo only
mad_subset_oecd <- read.csv("./input_data/maddison_subset_oecd.csv") # 1950-2008, 29 countries, 5-year CAGR, OECD members only
mad_subset_opec <- read.csv("./input_data/maddison_subset_opec.csv") # 1950-2008, 13 countries, 5-year CAGR, OPEC members only
# By excluded region
mad_subset_exeurope <- read.csv("./input_data/maddison_subset_exeurope.csv") # 1950-2008, 115 countries, 5-year CAGR, excluding Europe
mad_subset_exafrica <- read.csv("./input_data/maddison_subset_exafrica.csv") # 1950-2008, 89 countries, 5-year CAGR, excluding Sub-Saharan Africa
mad_subset_exeastasia <- read.csv("./input_data/maddison_subset_exeastasia.csv") # 1950-2008, 122 countries, 5-year CAGR, excluding East Asia
mad_subset_exlatam <- read.csv("./input_data/maddison_subset_exlatam.csv") # 1950-2008, 114 countries, 5-year CAGR, excluding Latin America
mad_subset_exanglo <- read.csv("./input_data/maddison_subset_exanglo.csv") # 1950-2008, 131 countries, 5-year CAGR, excluding Anglophone countries (USA, UK, Australia, Ireland, Canada)
mad_subset_exoecd <- read.csv("./input_data/maddison_subset_exoecd.csv") # 1950-2008, 108 countries, 5-year CAGR, excluding OECD members
mad_subset_exopec <- read.csv("./input_data/maddison_subset_exopec.csv") # 1950-2008, 124 countries, 5-year CAGR, excluding OPEC members
matrices <- create.dist.matrices(mad_subset5, add.trade=TRUE)
create.dist.matrices <- function (mad, add.trade=FALSE) {
# Get linguistic data from Ruhlen
ruhlen <- read.csv("./input_data/ruhlen.csv")
ruhlen <- rename(ruhlen, country_code = Country..ISO.Alpha.3.Code.) # Clean up column name for country code
# Get most spoken language for each country
ruhlen2 <- ruhlen %>%
group_by(country_code) %>%
arrange(desc(Population)) %>%
slice(1) %>% # In case there's more than one max language, just take the first one
ungroup
## Dealing with missing languages
# The real dominant languages of certain countries are missing
# E.g. For Australia, there's no Australian English and so the dominant language is (incorrectly) a small aboriginal language
# In this section, I manually substitute the phonemic data for some countries' languages with others that make sense
# E.g. Substitute Australia with the phonemic data from Britain's English
ruhlen2[which(ruhlen2$country_code == "AFG"), 12:739] <- ruhlen[which(ruhlen$ISO_A3 == "pbt-AFG"), 12:739] # Pashto (not Farsi) is the largest language in Afghanistan
ruhlen2[which(ruhlen2$country_code == "AUS"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "GBR"), 12:739] # Australia <- Britain
ruhlen2[which(ruhlen2$country_code == "ARG"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Argentina <- Colombia
ruhlen2[which(ruhlen2$country_code == "BOL"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Bolivia <- Colombia
ruhlen2[which(ruhlen2$country_code == "BWA"), 12:739] <- ruhlen[which(ruhlen$ISO_A3 == "tsn-ZAF"), 12:739] # Tswana is main language of Botswana (but categorized as a language of South Africa)
ruhlen2[which(ruhlen2$country_code == "CHE"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "DEU"), 12:739] # Switzerland <- Germany
ruhlen2[which(ruhlen2$country_code == "CHL"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Chile <- Colombia
ruhlen2[which(ruhlen2$country_code == "GTM"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "CRI"), 12:739] # Guatemala <- Costa Rica
ruhlen2[which(ruhlen2$country_code == "HND"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "CRI"), 12:739] # Honduras <- Costa Rica
ruhlen2[which(ruhlen2$country_code == "NIC"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "CRI"), 12:739] # Nicaragua <- Costa Rica
ruhlen2[which(ruhlen2$country_code == "NZL"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "GBR"), 12:739] # New Zealand <- Britain
ruhlen2[which(ruhlen2$country_code == "PER"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Peru <- Colombia
ruhlen2[which(ruhlen2$country_code == "PRY"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Paraguay <- Colombia
ruhlen2[which(ruhlen2$country_code == "SDN"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "EGY"), 12:739] # Sudan <- Egypt
ruhlen2[which(ruhlen2$country_code == "SLV"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "CRI"), 12:739] # El Salvador <- Costa Rica
ruhlen2[which(ruhlen2$country_code == "VEN"), 12:739] <- ruhlen2[which(ruhlen2$country_code == "COL"), 12:739] # Venezuela <- Colombia
# Get geograhic data for each country (longitude and latitude)
geo <- read.csv("./input_data/geo.csv")
# Get 2008 levels of income for each country
mad08 <- read.csv("./input_data/maddison_subset_2008gdp.csv")
if (add.trade) {
# Trade distance matrix - direct input method (i.e. data is already cleaned up in Excel)
trade.dist.matrix <- read.csv("./input_data/trade_distance_matrix.csv")
colnames(trade.dist.matrix)[1] <- "country_code"
}
# Harmonize all datasets by selecting country rows that they all have
full <- merge(ruhlen2, mad, by = "country_code")
full <- merge(full, geo, by = "country_code")
if (add.trade) {
full <- merge(full, trade.dist.matrix, by = "country_code") ############ Trade
}
full <- select(full, country_code)
ruhlen2 <- merge(full, ruhlen2, by = 'country_code')
mad2 <- merge(full, mad, by = 'country_code')
geo2 <- merge(full, geo, by = 'country_code')
mad08 <- merge(full, mad08, by = 'country_code')
if (add.trade) {
trade.dist.matrix <- merge(full, trade.dist.matrix, by = 'country_code') ############ Trade
}
identical(as.vector(mad2$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
identical(as.vector(geo2$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
if (add.trade) {
identical(as.vector(trade.dist.matrix$country_code), as.vector(ruhlen2$country_code)) # Check to make sure country codes match up exactly
}
if (add.trade) {
rownames(trade.dist.matrix) <- trade.dist.matrix[, 1]
trade.dist.matrix <- trade.dist.matrix[, -1]
trade.dist.matrix <- as.data.frame(t(trade.dist.matrix))
trade.dist.matrix$country_code <- rownames(trade.dist.matrix)
trade.dist.matrix <- merge(full, trade.dist.matrix, by = 'country_code')
rownames(trade.dist.matrix) <- trade.dist.matrix[, 1]
trade.dist.matrix <- trade.dist.matrix[, -1]
trade.dist.matrix <- as.dist(trade.dist.matrix)
}
# Create growth distance matrix based on Maddison income data using Euclidean distance
mad3 <- mad2[, 3:ncol(mad2)] # Extract the growth numbers
rownames(mad3) <- mad2$country_code # Rename rows using country codes
mad3 <- t(scale(t(mad3))) # Standardize growth rates within each country so that mean = 0, SD = 1
growth.dist.matrix <- vegdist(mad3, "euclidean")
# Create language distance matrix for Ruhlen linguistic data using Jaccard distance
ruhlen3 <- ruhlen2[, 12:ncol(ruhlen2)] # Extract phoneme presence/absence for all 728 phonemes
rownames(ruhlen3) <- ruhlen2$country_code # Rename rows to ISO country codes
lang.dist.matrix <- vegdist(ruhlen3, "jaccard", binary = TRUE)
# Create geographic distance matrix for geographic data using great circle distance
geo3 <- geo2[ , 2:3]
geo.dist.matrix <- rdist.earth(geo3, geo3)
rownames(geo.dist.matrix) <- geo2[, 1]
colnames(geo.dist.matrix) <- geo2[, 1]
geo.dist.matrix <- log(as.dist(geo.dist.matrix)) # Log distance
# Create distance matrix for 2008 levels of income
mad08_2 <- as.matrix(mad08$X2008) # Extract the income numbers
rownames(mad08_2) <- mad08$country_code # Rename rows using country codes
mad08.dist.matrix <- log(vegdist(mad08_2, "euclidean")) # Log distance
if (add.trade) {
output <- list(growth.dist.matrix, lang.dist.matrix, geo.dist.matrix, mad08.dist.matrix, trade.dist.matrix)
} else {
output <- list(growth.dist.matrix, lang.dist.matrix, geo.dist.matrix, mad08.dist.matrix)
}
return(output)
}
matrices <- create.dist.matrices(mad_subset5, add.trade=TRUE)
matrices.w.trade <- create.dist.matrices(mad_subset5, add.trade=TRUE)
ncf::mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[5]]), resamp = 9999)
ncf::partial.mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[2]], as.matrix(matrices.w.trade[[5]]), resamp = 9999)
ncf::partial.mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[2]]), as.matrix(matrices.w.trade[[5]]), resamp = 9999)
ncf::partial.mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[2]]), as.matrix(matrices.w.trade[[5]]), resamp = 9999)
matrices.w.trade <- create.dist.matrices(mad_subset5, add.trade=TRUE)
ncf::partial.mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[2]]), as.matrix(matrices.w.trade[[5]]), resamp = 9999)
ncf::partial.mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[2]]), as.matrix(matrices.w.trade[[3]]), as.matrix(matrices.w.trade[[4]]), as.matrix(matrices.w.trade[[5]]), resamp = 9999)
View(as.matrix(matrices.w.trade[[2]]))
complete.cases(as.matrix(matrices.w.trade[[2]]))
complete.cases(as.matrix(matrices.w.trade[[5]]))
which(is.na(as.matrix(matrices.w.trade[[5]])))
?n()
n(which(is.na(as.matrix(matrices.w.trade[[5]]))))
length(which(is.na(as.matrix(matrices.w.trade[[5]]))))
nrow(as.matrix(matrices.w.trade[[5]]))
ncol(as.matrix(matrices.w.trade[[5]]))
sapply(as.matrix(matrices.w.trade[[5]])), is.na)
sapply(as.matrix(matrices.w.trade[[5]]), is.na)
length(sapply(as.matrix(matrices.w.trade[[5]]), is.na))
length(as.matrix(matrices.w.trade[[5]]))
sumif(sapply(as.matrix(matrices.w.trade[[5]]), is.na), TRUE)
sum(sapply(as.matrix(matrices.w.trade[[5]]), is.na))
trade.modified <- as.matrix(matrices.w.trade[[5]])
upper.tri(trade.modified)
trade.modified[upper.tri(trade.modified)] <- NA
length(sapply(trade.modified, is.na))
sum(sapply(trade.modified, is.na))
View(trade.modified)
trade.modified <- as.matrix(matrices.w.trade[[5]])
trade.modified[upper.tri(trade.modified, diag=TRUE)] <- NA
View(trade.modified)
sum(sapply(trade.modified, is.na))
sum(sapply(trade.modified, !is.na))
which( !is.na(trade.modified), arr.ind=TRUE)
n(which( !is.na(trade.modified), arr.ind=TRUE))
length(which( !is.na(trade.modified), arr.ind=TRUE))
sum(sapply(trade.modified, is.na))
write.csv(as.matrix(matrices.w.trade[[5]]), "./output_tables/trade_dist_matrix2.csv")
trade.modified <- as.matrix(matrices.w.trade[[5]])
trade.modified[upper.tri(trade.modified, diag=TRUE)] <- NA
write.csv(trade.modified), "./output_tables/trade_dist_matrix2.csv")
write.csv(trade.modified), "./output_tables/trade_dist_matrix2.csv")
trade.modified <- as.matrix(matrices.w.trade[[5]])
trade.modified[upper.tri(trade.modified, diag=TRUE)] <- NA
View(trade.modified)
write.csv(trade.modified), "./output_tables/trade_dist_matrix2.csv")
write.csv(as.matrix(trade.modified)), "./output_tables/trade_dist_matrix2.csv")
write.csv(as.matrix(trade.modified), "./output_tables/trade_dist_matrix2.csv")
ncf::partial.mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[2]]), as.matrix(matrices.w.trade[[5]]), resamp = 99999)
ncf.output <- data.frame(matrix(nrow=4))
ncf.output <- ncf::partial.mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[2]]), as.matrix(matrices.w.trade[[5]]), resamp = 9999)
ncf.output$MantelR
ncf.output$MantelR[1]
ncf.output$p
ncf.output$p[1]
ncf1 <- paste0(formatC(round(ncf.output$MantelR[1], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[1], scientific=TRUE, digits = 2),")")
ncf1
ncf1 <- paste0(formatC(round(ncf.output$MantelR[1], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[1], scientific=TRUE, digits = 2),")")
ncf2 <- paste0(formatC(round(ncf.output$MantelR[2], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[2], scientific=TRUE, digits = 2),")")
ncf3 <- paste0(formatC(round(ncf.output$MantelR[3], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[3], scientific=TRUE, digits = 2),")")
ncf4 <- paste0(formatC(round(ncf.output$MantelR[4], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[4], scientific=TRUE, digits = 2),")")
ncf.formatted <- data.frame(ncf1, ncf2, ncf3, ncf4)
View(ncf.formatted)
colnames(ncf.formatted) <- c("Growth ~ Linguistic", "Growth ~ Trade", "Linguistic ~ Trade", "Growth ~ Linguistic | Trade")
View(trade.modified)
View(ncf.formatted)
xtable(ncf.formatted)
print(xtable(ncf.formatted), include.rownames=FALSE)
colnames(ncf.formatted) <- c("Growth-Linguistic", "Growth-Trade", "Linguistic-Trade", "Growth-Linguistic (Control: Trade)")
print(xtable(ncf.formatted), include.rownames=FALSE)
print(xtable(ncf.formatted), include.rownames=FALSE, align = c(rep("p{2.3cm}", 4)), type='latex', sanitize.text.function=identity)
print(xtable(ncf.formatted, align = c(rep("p{2.3cm}", 4)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
print(xtable(ncf.formatted, align = c(rep("p{2.3cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
print(xtable(ncf.formatted, align = c(rep("p{3cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
ncf.output <- ncf::partial.mantel.test(as.matrix(matrices.w.trade[[1]]), as.matrix(matrices.w.trade[[2]]), as.matrix(matrices.w.trade[[5]]), resamp = 99999)
ncf1 <- paste0(formatC(round(ncf.output$MantelR[1], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[1], scientific=TRUE, digits = 2),")")
ncf2 <- paste0(formatC(round(ncf.output$MantelR[2], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[2], scientific=TRUE, digits = 2),")")
ncf3 <- paste0(formatC(round(ncf.output$MantelR[3], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[3], scientific=TRUE, digits = 2),")")
ncf4 <- paste0(formatC(round(ncf.output$MantelR[4], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[4], scientific=TRUE, digits = 2),")")
ncf.formatted <- data.frame(ncf1, ncf2, ncf3, ncf4)
colnames(ncf.formatted) <- c("Growth-Linguistic", "Growth-Trade", "Linguistic-Trade", "Growth-Linguistic \\hspace{20 mm} (Control: Trade)")
print(xtable(ncf.formatted, align = c(rep("p{3cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
ncf1 <- paste0(formatC(round(ncf.output$MantelR[1], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[1], scientific=TRUE, digits = 2),")")
ncf2 <- paste0(formatC(round(ncf.output$MantelR[2], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[2], scientific=TRUE, digits = 2),")")
ncf3 <- paste0(formatC(round(ncf.output$MantelR[3], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[3], scientific=TRUE, digits = 2),")")
ncf4 <- paste0(formatC(round(ncf.output$MantelR[4], digits = 3), digits = 3, format = "f"), " (",format(ncf.output$p[4], scientific=TRUE, digits = 2),")")
ncf.formatted <- data.frame(ncf1, ncf2, ncf3, ncf4)
View(ncf.formatted)
colnames(ncf.formatted) <- c("Growth-Linguistic", "Growth-Trade", "Linguistic-Trade", "Growth-Linguistic \\hspace{20 mm} (Control: Trade)")
View(ncf.formatted)
print(xtable(ncf.formatted, align = c(rep("p{3cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
colnames(ncf.formatted) <- c("Growth-Linguistic", "Growth-Trade", "Linguistic-Trade", "Growth-Linguistic (Control: Trade)")
print(xtable(ncf.formatted, align = c(rep("p{3cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
colnames(ncf.formatted) <- c("(1) Growth-Linguistic", "(2) Growth-Trade", "(3) Linguistic-Trade", "(4) Growth-Linguistic (Control: Trade)")
print(xtable(ncf.formatted, align = c(rep("p{3cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
print(xtable(ncf.formatted, align = c(rep("p{3.1cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
print(xtable(ncf.formatted, align = c(rep("p{3.3cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
print(xtable(ncf.formatted, align = c(rep("p{3.5cm}", 5)), type='latex', sanitize.text.function=identity), include.rownames=FALSE)
matrices <- create.dist.matrices(mad_subset5)
correlogram1 <- mgram(matrices[[1]], matrices[[3]])
plot(correlogram1)
correlogram1 <- mgram(matrices[[1]], matrices[[3]], breaks=5)
plot(correlogram1)
correlogram1 <- mgram(matrices[[1]], matrices[[3]], nclass=5)
plot(correlogram1)
correlogram1 <- mgram(matrices[[1]], matrices[[3]], nclass=10)
plot(correlogram1)
View(matrices[[1]])
View(as.matrix(matrices[[1]]))
correlogram1
correlogram1 <- mgram(matrices[[1]], matrices[[3]], nclass=20)
plot(correlogram1)
correlogram1
max(as.matrixmatrices[[3]]))
max(as.matrixmatrices[[3]])
max(as.matrix(matrices[[3]]))
correlogram1 <- mgram(matrices[[1]], matrices[[3]], breaks=c(5, 6, 7, 8, 9))
plot(correlogram1)
correlogram1
correlogram1 <- mgram(matrices[[1]], matrices[[3]], breaks=c(1:9))
plot(correlogram1)
matrices <- create.dist.matrices(mad_subset5)
growth.dist.matrix <- matrices[[1]]
lang.dist.matrix <- matrices[[2]]
geo.dist.matrix <- matrices[[3]]
mad08.dist.matrix <- matrices[[4]]
growth.pairs <- convert.dist.matrix(growth.dist.matrix)
geo.pairs <- convert.dist.matrix(geo.dist.matrix)
growth.pairs2 <- growth.pairs
growth.pairs2 <- transmute(growth.pairs2, pairs = paste(as.character(country1), as.character(country2)), value)
geo.pairs2 <- geo.pairs
geo.pairs2 <- transmute(geo.pairs2, pairs = paste(as.character(country1), as.character(country2)), value)
merged.pairs <- merge(growth.pairs2, geo.pairs2, by = "pairs")
colnames(merged.pairs) <- c("pair", "growth", "geo")
ggplot(merged.pairs, aes(growth, geo)) +
geom_text(aes(label=pair)) +
geom_smooth(method = "lm", se = FALSE)
convert.dist.matrix <- function(dist.matrix) {
dist.matrix <- as.matrix(dist.matrix)
dist.matrix[upper.tri(dist.matrix, diag=TRUE)] <- NA
df <- melt(dist.matrix, varnames = c("country1", "country2"))
df <- na.omit(df)
return(df)
}
growth.pairs <- convert.dist.matrix(growth.dist.matrix)
geo.pairs <- convert.dist.matrix(geo.dist.matrix)
growth.pairs2 <- growth.pairs
growth.pairs2 <- transmute(growth.pairs2, pairs = paste(as.character(country1), as.character(country2)), value)
geo.pairs2 <- geo.pairs
geo.pairs2 <- transmute(geo.pairs2, pairs = paste(as.character(country1), as.character(country2)), value)
merged.pairs <- merge(growth.pairs2, geo.pairs2, by = "pairs")
colnames(merged.pairs) <- c("pair", "growth", "geo")
ggplot(merged.pairs, aes(growth, geo)) +
geom_text(aes(label=pair)) +
geom_smooth(method = "lm", se = FALSE)
View(merged.pairs)
View(arrange(merged.pairs, growth))
growth.pairs <- convert.dist.matrix(growth.dist.matrix)
lang.pairs <- convert.dist.matrix(lang.dist.matrix)
geo.pairs <- convert.dist.matrix(geo.dist.matrix)
growth.pairs2 <- growth.pairs
growth.dist.matrix <- matrices[[1]]
lang.dist.matrix <- matrices[[2]]
geo.dist.matrix <- matrices[[3]]
mad08.dist.matrix <- matrices[[4]]
# Compare growth and geo
growth.pairs <- convert.dist.matrix(growth.dist.matrix)
lang.pairs <- convert.dist.matrix(lang.dist.matrix)
geo.pairs <- convert.dist.matrix(geo.dist.matrix)
growth.pairs2 <- growth.pairs
growth.pairs2 <- transmute(growth.pairs2, pairs = paste(as.character(country1), as.character(country2)), value)
lang.pairs2 <- lang.pairs
lang.pairs2 <- transmute(lang.pairs2, pairs = paste(as.character(country1), as.character(country2)), value)
geo.pairs2 <- geo.pairs
geo.pairs2 <- transmute(geo.pairs2, pairs = paste(as.character(country1), as.character(country2)), value)
merged.pairs <- merge(growth.pairs2, lang.pairs2, geo.pairs2, by = "pairs")
colnames(merged.pairs) <- c("pair", "growth", "lang", "geo")
View(arrange(merged.pairs, growth))
View(lang.pairs2)
merged.pairs <- merge(growth.pairs2, lang.pairs2, by = "pairs")
merged.pairs <- merge(merged.pairs, geo.pairs2, by = "pairs")
colnames(merged.pairs) <- c("pair", "growth", "lang", "geo")
View(arrange(merged.pairs, growth))