Gini values are not corresponding with Lorenz Curve area

Question

I'm using Gini coefficient and Lorenz Curve plots to show the accumulation of beneficiaries in ecosystem services (ES) supply points, in R. I classify ES into three categories and calculate Gini and Lorenz curve for each category. However, the highest value doesn't correspond with the biggest area in the plot. Why is this happening?

Here you can download the dataframe

And here is the script:

library(ggplot2)
library(gglorenz)
library(ineq)

occ_hubs <- read.csv2('occ_hubs.csv')

#calculate GINI per categories
Category <- 1
Gini_value <- 1
j <- 1
for (i in unique(occ_hubs$categoria)){
  subset <- subset(occ_hubs, categoria == i)
  gini_cat <- Gini(subset$id_hub, subset$Freq)
  Category[j] <- i
  Gini_value[j] <- gini_cat
  j <- j+1
  print(c(i, gini_cat))
}
ginis_cat <- as.data.frame(Category)
ginis_SE$Gini <- Gini_value

#Plot Lorenz Curve by cateogries
ggplot(occ_hubs, aes(x=Freq, color=categoria)) +
  stat_lorenz(desc = TRUE, size=1) +
  coord_fixed() +
  geom_abline(linetype = "dashed") +
  theme_minimal() +
  hrbrthemes::scale_x_percent() +
  hrbrthemes::scale_y_percent() +
  hrbrthemes::theme_ipsum() +
  labs(x = "Cumulative percentage of hubs",
       y = "Cumulative Percentage of total flow",
       legend.title='')+
  theme(legend.title=element_blank())

Thank you.

score 0 · Answer 1 · answered May 01 '20 at 08:07

I found my mistake. My df has not two parameters for Gini coefficient since 'id_hub' is always 1 (it's a factor, indeed), so 'x' must be 'Freq' and 'parameter' left without specification:

library(ggplot2)
library(gglorenz)
library(ineq)

occ_hubs <- read.csv2('occ_hubs.csv')

#calculate GINI per categories
Category <- 1
Gini_value <- 1
j <- 1
for (i in unique(occ_hubs$categoria)){
  subset <- subset(occ_hubs, categoria == i)
  gini_cat <- Gini(subset$Freq)
  Category[j] <- i
  Gini_value[j] <- gini_cat
  j <- j+1
  print(c(i, gini_cat))
}
ginis_cat <- as.data.frame(Category)
ginis_SE$Gini <- Gini_value

#Plot Lorenz Curve by cateogries
ggplot(occ_hubs, aes(x=Freq, color=categoria)) +
  stat_lorenz(desc = TRUE, size=1) +
  coord_fixed() +
  geom_abline(linetype = "dashed") +
  theme_minimal() +
  hrbrthemes::scale_x_percent() +
  hrbrthemes::scale_y_percent() +
  hrbrthemes::theme_ipsum() +
  labs(x = "Cumulative percentage of hubs",
       y = "Cumulative Percentage of total flow",
       legend.title='')+
  theme(legend.title=element_blank())

Gini values are not corresponding with Lorenz Curve area

1 Answers1