corp13: Top 1000 corporations in Denmark in 2013
In antongrau/eliter: Contains the Danish Elite Network for 2012 and 2017 with analytical functions.
corp13 R Documentation
Top 1000 corporations in Denmark in 2013
Description
This dataset was used for the article: "Who listens to the top? Integration of the largest corporations across sectoral networks" published in Acta Sociologica and written by Anton Grau Larsen and Christoph Ellersgaard in 2017.
The example is the code that replicates the analysis in that article.
Examples
# Who listens to the top -----
library(eliter)
library(Matrix)
library(RColorBrewer)
library(ggthemes)
library(car)
data(corp13)
# Active variables ------
active <- data.frame(Turnover = rank(corp13$OMSÆTNING.10, na.last = "keep"),
Employees = rank(corp13$ANSATTE.10, na.last = "keep"),
Equity = rank(corp13$EGENKAPITAL.10, na.last = "keep"),
Dominating = corp13$dominerende_num,
Finance = corp13$finance,
Coop = corp13$coop,
Global500 = corp13$Global,
Books = corp13$books,
"Radio and Tv" = corp13$radio_tv
)
rownames(active) <- corp13$ORG_NAVN
colSums(sapply(active, is.na))
active.complete <- na.omit(active)
num.mem <- corp13[, grep("Memberships",colnames(corp13))]
# Table 2: Network statistics for each sectorial network ------
# § A table, divided by sector, with number of affiliations, individuals, largest component, diameter of the largest component
describe.sector <- function(x){
out <- list()
out$"Affiliations" <- length(table(droplevels(x$AFFILIATION)))
out$"Postions" <- nrow(x)
out$"Individuals" <- length(table(droplevels(x$NAME)))
net.sector <- eliter::elite.network(droplevels(x), result = "affil")
net.sector.com <- largest.component(net.sector, cut.off = 0.001)
out$"Largest component" <- vcount(net.sector.com)
out$"Component diameter" <- diameter(net.sector.com, weights = NA, directed = FALSE)
out$"Highest degree" <- max(degree(net.sector, loops = FALSE, mode = "out"))
out
}
table.2 <- t(sapply(sector.rels, describe.sector))
#write.csv(table.2, file = "output_corporate/Table_2_Network_stats_per_sector.csv")
# Table 3: Mean number of memberships per sector -----
table.3.list <- list()
table.3.list$Finance <- aggregate(num.mem, by = list(corp13$finance), mean, na.rm = TRUE)[2, -1]
table.3.list$Dominant <- aggregate(num.mem, by = list(corp13$dominerende_num), mean, na.rm = TRUE)[2, -1]
table.3.list$Coop <- aggregate(num.mem, by = list(corp13$coop), mean, na.rm = TRUE)[2, -1]
table.3.list$"Global 500" <- aggregate(num.mem, by = list(corp13$Global), mean, na.rm = TRUE)[2, -1]
table.3.list$"All" <- apply(num.mem, 2, mean, na.rm = TRUE)
table.3b.list <- list()
table.3b.list$Finance <- aggregate(num.mem, by = list(finance), sd, na.rm = TRUE)[2, -1]
table.3b.list$Dominant <- aggregate(num.mem, by = list(corp13$dominerende_num), sd, na.rm = TRUE)[2, -1]
table.3b.list$Coop <- aggregate(num.mem, by = list(corp13$coop), sd, na.rm = TRUE)[2, -1]
table.3b.list$"Global 500" <- aggregate(num.mem, by = list(corp13$Global), sd, na.rm = TRUE)[2, -1]
table.3b.list$"All" <- apply(num.mem, 2, sd, na.rm = TRUE)
table.3b <- t(do.call(what = rbind, args = table.3b.list))
table.3b <- round(table.3b, 1)
rownames(table.3b) <- sapply(str_split(rownames(table.3b), " : "), tail, 1)
tab.3.n <- c(sum(corp13$finance, na.rm = TRUE),
sum(corp13$dominerende_num),
sum(corp13$coop),
sum(corp13$Global),
nrow(num.mem)
)
table.3 <- t(do.call(what = rbind, args = table.3.list))
table.3 <- round(table.3, 1)
rownames(table.3) <- sapply(str_split(rownames(table.3), " : "), tail, 1)
table.3[] <- paste(table.3[], " (", table.3b[], ")", sep = "")
table.3 <- rbind(table.3, N = tab.3.n)
#write.csv(table.3, file = "output_corporate/Table_3_Mean_number_of_memberships_per_sector.csv")
# Numbers for big linkers paragraph ----
net.corp.all <- elite.network(sector.rels$Corporations)
net.corp.com <- largest.component(net.corp.all, cut.off = 0)
table(V(net.corp.com)$memberships)
connectors <- V(net.corp.com)$name[V(net.corp.com)$memberships > 1]
chairmen <- unique(sector.rels$Corporation$NAME[sector.rels$Corporation$ROLE == "Chairman"])
ceo <- unique(sector.rels$Corporation$NAME[sector.rels$Corporation$ROLE == "Chief executive"])
nonordinary <- unique(sector.rels$Corporation$NAME[sector.rels$Corporation$ROLE %in% c("Chief executive", "Chairman", "Vice chairman", "Executive")])
length(connectors)
sum(connectors %in% chairmen)
sum(connectors %in% ceo)
sum(connectors %in% nonordinary)
ordinary <- connectors[(connectors %in% nonordinary) == FALSE]
length(ordinary)
chairmen.all <- V(net.corp.all)$name %in% chairmen
ceo.all <- V(net.corp.all)$name %in% ceo
ordinary.all <- (V(net.corp.all)$name %in% nonordinary) == FALSE
table(ordinary.all)
table(chairmen.all)
rels.corp <- sector.rels$Corporation
rels.corp <- rels.corp[-which((rels.corp$NAME %in% nonordinary) == FALSE),]
net.corp.org.red <- elite.network(rels.corp, result = "affil")
net.corp.org.red <- largest.component(net.corp.org.red, cut.off = 0)
net.com.org <- elite.network(sector.rels$Corporations, result = "affil")
net.com.org <- largest.component(net.com.org, cut.off = 0)
vcount(net.corp.org.red) - vcount(net.com.org)
# Correlation lines -----
changes <- list()
changes$vline <- geom_vline(xintercept = nrow(corp13) - 250, linetype = "dashed", color = "grey30")
changes$ylab <- ylab("Memberships")
d.p <- data.frame(num.mem, omsætning = rank(corp13$OMSÆTNING, na.last = "keep"), navn = corp13$ORG_NAVN, check.names = FALSE)
mdp <- melt(d.p, id.vars = c("navn", "omsætning"))
p.cor.line.turnover <- ggplot(data = mdp, aes(x = omsætning, y = value)) + geom_point(shape = 21, alpha = 0.3, size = 1.2, fill = "whitesmoke") + facet_wrap(~variable, scales = "free_y") + geom_smooth(color = "red2", method = "loess") + theme_bw()
p.cor.line.turnover <- p.cor.line.turnover + changes + xlab("Rank by turnover")
p.cor.line.turnover <- p.cor.line.turnover + theme_tufte() + theme(strip.text = element_text(size = 12))
# pdf(file = "output_corporate/Figure_2_Correlation_smooths_by_sector.pdf", height = 10, width = 12)
# p.cor.line.turnover
# dev.off()
# Regression -----
reg <- as.matrix(active)
num.mem <- as.matrix(num.mem)
models.memberships <- lm(num.mem ~ reg)
d <- data.frame(num.mem, reg)
m1 <- lm(Memberships...Corporations ~ Turnover + Employees + Equity + Dominating + Finance + Coop + Global500 + Books + Radio.and.Tv, data = d)
vif.val <- car::vif(m1)
reg.stats <- coef(summary(models.memberships))
for (i in seq_along(reg.stats)) colnames(reg.stats[[i]]) <- paste(names(reg.stats)[i],colnames(reg.stats[[i]]))
sum.mod <- summary(models.memberships)
r.squared <- vector(length = length(sum.mod))
for (i in seq_along(sum.mod)) r.squared[i] <- sum.mod[[i]]$adj.r.squared
r.squared <- round(r.squared, 2)
reg.stats <- do.call(cbind, reg.stats)
reg.estimate <- reg.stats[, grep("Estimate", colnames(reg.stats))]
reg.pvalue <- reg.stats[, grep("Pr(>|t|)", colnames(reg.stats))]
reg.error <- reg.stats[, grep("Error", colnames(reg.stats))]
reg.stars <- apply(reg.pvalue, 2, symnum, na = FALSE, cutpoints = c(0, 0.001, 0.01, 0.05, 0.1, 1), symbols = c("***", "**", "*", ".", " "))
est.min <- format(round(reg.estimate - reg.error, 4), scientific = FALSE)
est.max <- format(round(reg.estimate + reg.error, 4), scientific = FALSE)
est.error <- paste("[", est.min, ":", est.max, "]", sep = "")
est.error <- format(round(reg.error, 5))
reg.res <- reg.estimate
reg.res[] <- paste(format(round(reg.estimate, 5), scientific = FALSE), " (", reg.stars, ")", est.error, sep = "")
reg.res <- rbind(reg.res, R.squared = r.squared)
reg.range <- apply(reg, 2, range, na.rm = TRUE)
reg.range <- paste("[", reg.range[1,], ":", reg.range[2,],"]", sep = "")
colnames(reg.res) <- names(sector.rels)
reg.res <- cbind(Range = c("", reg.range, ""), Vif = c("", round(vif.val, 2), ""), reg.res)
reg.res <- rbind(reg.res, mean = c("", "", round(apply(num.mem, 2, mean), 2)))
reg.res <- rbind(reg.res, max = c("", "", round(apply(num.mem, 2, max), 2)))
reg.res
#write.csv(reg.res, file = "output_corporate/table.4_Regressions.csv")
###############################################################################
# Figure 1 - The corporate network
market.leader <- corp13$dominerende
levels(market.leader) <- c("Incumbent", "Dominated")
net.corp <- elite.network(sector.rels$Corporations, result = "affil")
lay <- layout_with_fr(net.corp, weights = E(net.corp)$weight^2, grid = "nogrid") * -1
p <- graph.plot(net.corp, lay, vertex.fill = market.leader, vertex.size = num.mem[,9], edge.alpha = 0.1, edge.color = "black", edge.size = 0.5)
p <- p + scale_fill_manual(values = c("red2", "white"), name = "Market position") + scale_size_continuous(range = c(1, 6), name = "Non-corporate \nmemberships")
p
# pdf(file = "output_corporate/Figure.1_corporate_network.pdf", height = 11, width = 12)
# p
# dev.off()
# Appendix -----
deg <- degree(net.corp)
com <- V(net.corp)$name %in% V(net.com.org)$name
turn <- corp13$OMSÆTNING.10
mem <- num.mem[,9]
com <- factor(com, labels = c("No", "Yes"))
rd <- data.frame(Name = V(net.corp)$name, "Turnover" = turn, "Non-corporate memberships" = mem, "Degree" = deg, "In component" = com, check.names = FALSE)
rd.com <- rd[com == "Yes",]
rd.com <- rd.com[order(rd.com$Turnover, decreasing = TRUE),]
rd.notcom <- rd[com == "No",]
rd.notcom <- rd.notcom[order(rd.notcom$Turnover, decreasing = TRUE),]
# write.csv(head(rd.com, 25), file = "output_corporate/Appendix table_com.csv", row.names = FALSE)
# write.csv(head(rd.notcom, 25), file = "output_corporate/Appendix table_notcom.csv", row.names = FALSE)
antongrau/eliter documentation built on March 2, 2024, 8:05 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
| corp13 | R Documentation |
Top 1000 corporations in Denmark in 2013
Description
This dataset was used for the article: "Who listens to the top? Integration of the largest corporations across sectoral networks" published in Acta Sociologica and written by Anton Grau Larsen and Christoph Ellersgaard in 2017. The example is the code that replicates the analysis in that article.
Examples
# Who listens to the top -----
library(eliter)
library(Matrix)
library(RColorBrewer)
library(ggthemes)
library(car)
data(corp13)
# Active variables ------
active <- data.frame(Turnover = rank(corp13$OMSÆTNING.10, na.last = "keep"),
Employees = rank(corp13$ANSATTE.10, na.last = "keep"),
Equity = rank(corp13$EGENKAPITAL.10, na.last = "keep"),
Dominating = corp13$dominerende_num,
Finance = corp13$finance,
Coop = corp13$coop,
Global500 = corp13$Global,
Books = corp13$books,
"Radio and Tv" = corp13$radio_tv
)
rownames(active) <- corp13$ORG_NAVN
colSums(sapply(active, is.na))
active.complete <- na.omit(active)
num.mem <- corp13[, grep("Memberships",colnames(corp13))]
# Table 2: Network statistics for each sectorial network ------
# § A table, divided by sector, with number of affiliations, individuals, largest component, diameter of the largest component
describe.sector <- function(x){
out <- list()
out$"Affiliations" <- length(table(droplevels(x$AFFILIATION)))
out$"Postions" <- nrow(x)
out$"Individuals" <- length(table(droplevels(x$NAME)))
net.sector <- eliter::elite.network(droplevels(x), result = "affil")
net.sector.com <- largest.component(net.sector, cut.off = 0.001)
out$"Largest component" <- vcount(net.sector.com)
out$"Component diameter" <- diameter(net.sector.com, weights = NA, directed = FALSE)
out$"Highest degree" <- max(degree(net.sector, loops = FALSE, mode = "out"))
out
}
table.2 <- t(sapply(sector.rels, describe.sector))
#write.csv(table.2, file = "output_corporate/Table_2_Network_stats_per_sector.csv")
# Table 3: Mean number of memberships per sector -----
table.3.list <- list()
table.3.list$Finance <- aggregate(num.mem, by = list(corp13$finance), mean, na.rm = TRUE)[2, -1]
table.3.list$Dominant <- aggregate(num.mem, by = list(corp13$dominerende_num), mean, na.rm = TRUE)[2, -1]
table.3.list$Coop <- aggregate(num.mem, by = list(corp13$coop), mean, na.rm = TRUE)[2, -1]
table.3.list$"Global 500" <- aggregate(num.mem, by = list(corp13$Global), mean, na.rm = TRUE)[2, -1]
table.3.list$"All" <- apply(num.mem, 2, mean, na.rm = TRUE)
table.3b.list <- list()
table.3b.list$Finance <- aggregate(num.mem, by = list(finance), sd, na.rm = TRUE)[2, -1]
table.3b.list$Dominant <- aggregate(num.mem, by = list(corp13$dominerende_num), sd, na.rm = TRUE)[2, -1]
table.3b.list$Coop <- aggregate(num.mem, by = list(corp13$coop), sd, na.rm = TRUE)[2, -1]
table.3b.list$"Global 500" <- aggregate(num.mem, by = list(corp13$Global), sd, na.rm = TRUE)[2, -1]
table.3b.list$"All" <- apply(num.mem, 2, sd, na.rm = TRUE)
table.3b <- t(do.call(what = rbind, args = table.3b.list))
table.3b <- round(table.3b, 1)
rownames(table.3b) <- sapply(str_split(rownames(table.3b), " : "), tail, 1)
tab.3.n <- c(sum(corp13$finance, na.rm = TRUE),
sum(corp13$dominerende_num),
sum(corp13$coop),
sum(corp13$Global),
nrow(num.mem)
)
table.3 <- t(do.call(what = rbind, args = table.3.list))
table.3 <- round(table.3, 1)
rownames(table.3) <- sapply(str_split(rownames(table.3), " : "), tail, 1)
table.3[] <- paste(table.3[], " (", table.3b[], ")", sep = "")
table.3 <- rbind(table.3, N = tab.3.n)
#write.csv(table.3, file = "output_corporate/Table_3_Mean_number_of_memberships_per_sector.csv")
# Numbers for big linkers paragraph ----
net.corp.all <- elite.network(sector.rels$Corporations)
net.corp.com <- largest.component(net.corp.all, cut.off = 0)
table(V(net.corp.com)$memberships)
connectors <- V(net.corp.com)$name[V(net.corp.com)$memberships > 1]
chairmen <- unique(sector.rels$Corporation$NAME[sector.rels$Corporation$ROLE == "Chairman"])
ceo <- unique(sector.rels$Corporation$NAME[sector.rels$Corporation$ROLE == "Chief executive"])
nonordinary <- unique(sector.rels$Corporation$NAME[sector.rels$Corporation$ROLE %in% c("Chief executive", "Chairman", "Vice chairman", "Executive")])
length(connectors)
sum(connectors %in% chairmen)
sum(connectors %in% ceo)
sum(connectors %in% nonordinary)
ordinary <- connectors[(connectors %in% nonordinary) == FALSE]
length(ordinary)
chairmen.all <- V(net.corp.all)$name %in% chairmen
ceo.all <- V(net.corp.all)$name %in% ceo
ordinary.all <- (V(net.corp.all)$name %in% nonordinary) == FALSE
table(ordinary.all)
table(chairmen.all)
rels.corp <- sector.rels$Corporation
rels.corp <- rels.corp[-which((rels.corp$NAME %in% nonordinary) == FALSE),]
net.corp.org.red <- elite.network(rels.corp, result = "affil")
net.corp.org.red <- largest.component(net.corp.org.red, cut.off = 0)
net.com.org <- elite.network(sector.rels$Corporations, result = "affil")
net.com.org <- largest.component(net.com.org, cut.off = 0)
vcount(net.corp.org.red) - vcount(net.com.org)
# Correlation lines -----
changes <- list()
changes$vline <- geom_vline(xintercept = nrow(corp13) - 250, linetype = "dashed", color = "grey30")
changes$ylab <- ylab("Memberships")
d.p <- data.frame(num.mem, omsætning = rank(corp13$OMSÆTNING, na.last = "keep"), navn = corp13$ORG_NAVN, check.names = FALSE)
mdp <- melt(d.p, id.vars = c("navn", "omsætning"))
p.cor.line.turnover <- ggplot(data = mdp, aes(x = omsætning, y = value)) + geom_point(shape = 21, alpha = 0.3, size = 1.2, fill = "whitesmoke") + facet_wrap(~variable, scales = "free_y") + geom_smooth(color = "red2", method = "loess") + theme_bw()
p.cor.line.turnover <- p.cor.line.turnover + changes + xlab("Rank by turnover")
p.cor.line.turnover <- p.cor.line.turnover + theme_tufte() + theme(strip.text = element_text(size = 12))
# pdf(file = "output_corporate/Figure_2_Correlation_smooths_by_sector.pdf", height = 10, width = 12)
# p.cor.line.turnover
# dev.off()
# Regression -----
reg <- as.matrix(active)
num.mem <- as.matrix(num.mem)
models.memberships <- lm(num.mem ~ reg)
d <- data.frame(num.mem, reg)
m1 <- lm(Memberships...Corporations ~ Turnover + Employees + Equity + Dominating + Finance + Coop + Global500 + Books + Radio.and.Tv, data = d)
vif.val <- car::vif(m1)
reg.stats <- coef(summary(models.memberships))
for (i in seq_along(reg.stats)) colnames(reg.stats[[i]]) <- paste(names(reg.stats)[i],colnames(reg.stats[[i]]))
sum.mod <- summary(models.memberships)
r.squared <- vector(length = length(sum.mod))
for (i in seq_along(sum.mod)) r.squared[i] <- sum.mod[[i]]$adj.r.squared
r.squared <- round(r.squared, 2)
reg.stats <- do.call(cbind, reg.stats)
reg.estimate <- reg.stats[, grep("Estimate", colnames(reg.stats))]
reg.pvalue <- reg.stats[, grep("Pr(>|t|)", colnames(reg.stats))]
reg.error <- reg.stats[, grep("Error", colnames(reg.stats))]
reg.stars <- apply(reg.pvalue, 2, symnum, na = FALSE, cutpoints = c(0, 0.001, 0.01, 0.05, 0.1, 1), symbols = c("***", "**", "*", ".", " "))
est.min <- format(round(reg.estimate - reg.error, 4), scientific = FALSE)
est.max <- format(round(reg.estimate + reg.error, 4), scientific = FALSE)
est.error <- paste("[", est.min, ":", est.max, "]", sep = "")
est.error <- format(round(reg.error, 5))
reg.res <- reg.estimate
reg.res[] <- paste(format(round(reg.estimate, 5), scientific = FALSE), " (", reg.stars, ")", est.error, sep = "")
reg.res <- rbind(reg.res, R.squared = r.squared)
reg.range <- apply(reg, 2, range, na.rm = TRUE)
reg.range <- paste("[", reg.range[1,], ":", reg.range[2,],"]", sep = "")
colnames(reg.res) <- names(sector.rels)
reg.res <- cbind(Range = c("", reg.range, ""), Vif = c("", round(vif.val, 2), ""), reg.res)
reg.res <- rbind(reg.res, mean = c("", "", round(apply(num.mem, 2, mean), 2)))
reg.res <- rbind(reg.res, max = c("", "", round(apply(num.mem, 2, max), 2)))
reg.res
#write.csv(reg.res, file = "output_corporate/table.4_Regressions.csv")
###############################################################################
# Figure 1 - The corporate network
market.leader <- corp13$dominerende
levels(market.leader) <- c("Incumbent", "Dominated")
net.corp <- elite.network(sector.rels$Corporations, result = "affil")
lay <- layout_with_fr(net.corp, weights = E(net.corp)$weight^2, grid = "nogrid") * -1
p <- graph.plot(net.corp, lay, vertex.fill = market.leader, vertex.size = num.mem[,9], edge.alpha = 0.1, edge.color = "black", edge.size = 0.5)
p <- p + scale_fill_manual(values = c("red2", "white"), name = "Market position") + scale_size_continuous(range = c(1, 6), name = "Non-corporate \nmemberships")
p
# pdf(file = "output_corporate/Figure.1_corporate_network.pdf", height = 11, width = 12)
# p
# dev.off()
# Appendix -----
deg <- degree(net.corp)
com <- V(net.corp)$name %in% V(net.com.org)$name
turn <- corp13$OMSÆTNING.10
mem <- num.mem[,9]
com <- factor(com, labels = c("No", "Yes"))
rd <- data.frame(Name = V(net.corp)$name, "Turnover" = turn, "Non-corporate memberships" = mem, "Degree" = deg, "In component" = com, check.names = FALSE)
rd.com <- rd[com == "Yes",]
rd.com <- rd.com[order(rd.com$Turnover, decreasing = TRUE),]
rd.notcom <- rd[com == "No",]
rd.notcom <- rd.notcom[order(rd.notcom$Turnover, decreasing = TRUE),]
# write.csv(head(rd.com, 25), file = "output_corporate/Appendix table_com.csv", row.names = FALSE)
# write.csv(head(rd.notcom, 25), file = "output_corporate/Appendix table_notcom.csv", row.names = FALSE)
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.