Grant Data: 500 Households that get Social Grant in a certain region, the data comes from Social and Economic Survey. Download the dummy data: https://drive.google.com/file/d/1TpmoTgUmjPhHuqv3yCEqqEK3VIMUsbuQ/view?usp=sharing

grant <- read.csv("grant.csv")

#PFOODEXP: Proportion of Food Expenditure to Total Expenditure
#HH_Income: Household Income ($)
#HH_FOOD: Household Food Expenditure ($)
#HH_Loc: Household Location (0: Rural, 1: Urban)
#Educ_H: Household Head Education (Year)
#HH_Size: Household Family Member
#Gender_H: Household Head Gender (0: Female, 1: Male )
#Age_H: Household Head Age

#Remove 1st column
grant[1] <- NULL  
#change variable type to factor
grant$Gender_H <- as.factor(grant$Gender_H) 
grant$HH_Loc <- as.factor(grant$HH_Loc)

head(grant)
##   PFOODEXP  HH_Food HH_Loc Gender_H Age_H Educ_H HH_Size HH_Income
## 1 73.16933 675.1766      1        0    32     22       4  922.7590
## 2 69.77417 596.2286      1        1    27     22      13  854.5119
## 3 63.31992 585.4946      0        1    42     15       8  924.6610
## 4 71.77146 559.9470      0        1    26     22       5  780.1806
## 5 52.36861 554.8954      1        0    32      9       3 1059.5954
## 6 79.74131 551.4373      0        1    21     15       5  691.5328
summary(grant)
##     PFOODEXP        HH_Food      HH_Loc  Gender_H     Age_H      
##  Min.   :29.12   Min.   :165.9   0:380   0:167    Min.   :15.00  
##  1st Qu.:56.43   1st Qu.:189.0   1:120   1:333    1st Qu.:25.00  
##  Median :65.35   Median :218.3                    Median :34.00  
##  Mean   :64.74   Mean   :251.1                    Mean   :34.16  
##  3rd Qu.:75.59   3rd Qu.:286.6                    3rd Qu.:42.00  
##  Max.   :86.98   Max.   :675.2                    Max.   :80.00  
##      Educ_H         HH_Size         HH_Income     
##  Min.   : 3.00   Min.   : 1.000   Min.   : 198.6  
##  1st Qu.: 6.00   1st Qu.: 4.000   1st Qu.: 280.7  
##  Median :15.00   Median : 5.000   Median : 351.0  
##  Mean   :12.57   Mean   : 5.392   Mean   : 406.6  
##  3rd Qu.:16.00   3rd Qu.: 6.000   3rd Qu.: 489.0  
##  Max.   :23.00   Max.   :20.000   Max.   :1059.6
#Scatter Plot with Linear Line

library(ggplot2)
library(ggthemes)

ggplot(data=grant, aes(HH_Income,PFOODEXP, colour = Gender_H, size = HH_Size)) + 
  geom_point(alpha=0.8) + geom_smooth(method = "lm", se=FALSE) +
  ylab("% Food Expenditure")+ xlab("Household Income per Month ($)") +
  guides(color = guide_legend(override.aes = list(size=5, linetype = c(0,0)), title = "HH Gender"), 
         size = guide_legend(override.aes = list(linetype = c(0,0)), title = "H Size")) +
  scale_size_continuous(range = c(1, 8), breaks = c(1, 2, 4, 8))+
  scale_color_manual(labels = c("female","male"), values = c("hotpink","deepskyblue"))+
  labs( col = "Gender") +
  ggtitle("Scatterplot of Percentage of Food Expenditure vs Household Income per Month ($)") +
  scale_x_log10()+
  theme_bw()+
  theme(plot.title = element_text(size=10, face= "bold"))

#Scatter Plot with LOESS

library(ggplot2)
library(ggthemes)

ggplot(data=grant, aes(HH_Income,PFOODEXP, shape = Gender_H, colour = Gender_H, size = HH_Size)) + 
  geom_point(alpha=0.8) + geom_smooth(method = "loess") +
  ylab("% Food Expenditure")+ xlab("Household Income per month ($)") +
  guides(colour = FALSE, 
         size = FALSE,
         shape = guide_legend(override.aes = 
                 list(size=5, linetype = c(0,0), 
                      colour = c("azure4","gold")), title = "HH Gender")) +
  scale_size_continuous(range = c(1, 8), breaks = c(1, 2, 4, 8))+
  scale_shape_manual(labels = c("female","male"), values = c("f","m"))+
  scale_color_manual(labels = c("female","male"), values = c("azure4","gold"))+
  ggtitle("Scatterplot of Percentage of Food Expenditure vs Household Income per Month ($)") +
  scale_x_log10()+
  theme_bw()+
  theme(plot.title = element_text(size=10, face= "bold"))

Scatter Plot with LOESS: One of advantage of LOESS method, it doesn’t require the specification of a function to fit a model to all of the data in the sample. One of disadvantage, it doesn’t generate a regression function that is easily represented by mathematical formula.

#Hexbin Plot

library(hexbin)
library(RColorBrewer)

# Create data
  y<-grant%>% pull("HH_Food")
  x<-grant%>% pull("HH_Income")
  
# Make the plot
  bin<-hexbin(x, y)
  rf=colorRampPalette(rev(brewer.pal(10,'Spectral')))
  
  hexbinplot(y~x, data=bin, main="Income vs Food Expenditure",
             colramp=rf, trans=log, inv=exp,mincnt=1, maxcnt=70,
             ylab="food expenditure ($)",
             xlab="income ($)", cex.label=0.7)

Hexbin Plot: Scatterplots can get very hard to interpret when displaying large datasets, as points inevitably overplot. Hexbinplot helps discerning the data individually. Code source: www.everydayanalytics.ca

#Density Plot

qplot(HH_Income,data = grant, geom="density", fill = HH_Loc ,alpha=I(.5), 
      ylab="Density",
      xlab= "Household Income($)", 
      main = "Distribution of Household (HH) Income per Month by Household Location") + 
      scale_fill_manual(labels = c("rural","urban"), values = c("tomato","mediumspringgreen"))+
      labs( fill = "HH Location") + geom_density(alpha= 0.2,aes(HH_Income), colour = "grey85")+
      theme_minimal() + theme(plot.title = element_text(size=10, face= "bold"))

Income distribution is often right skewed, this shows income inequality. The hypothesized reasons are differences in talents, skills, and opportunities. It is not suprising, houshold income distribution in urban area is more skewed than in rural area.