2  Demographic data from census 2019

3 Age pyramid

4 Population Density

     District Population Area(km^2) Population Density
1           1     161975       7.72         20981.2176
2          10     239944       5.72         41948.2517
3          11     212448       5.14         41332.2957
4          12     489271      52.74          9277.0383
5           3     199914       4.92         40632.9268
6           4     174448       4.18         41733.9713
7           5     175828       4.27         41177.5176
8           6     261785       7.14         36664.5658
9           7     343530      35.69          9625.3853
10          8     403282      19.11         21103.1920
11 binh chanh     508490     252.56          2013.3434
12   binh tan     622294      52.02         11962.5913
13 binh thanh     515770      20.78         24820.5005
14    can gio      78240     704.45           111.0654
15     cu chi     419260     434.77           964.3260
16     go vap     664746      19.73         33692.1439
17    hoc mon     417053     109.17          3820.2162
18     nha be     119588     100.43          1190.7597
19  phu nhuan     182878       4.88         37475.0000
20   tan binh     395784      22.43         17645.2965
21    tan phu     516513      15.97         32342.7051
22    thu duc     965772     211.56          4565.0028

4.1 Social Variable

Ownership of dwelling

5 Economic variable

The wealth asset index was constructed by principal components analysis (PCA) using information on the ownership of dwelling, water supply, air condition, computer, washer, refrigerator, tv, radio, electricity, hot water, phone that are related to the household’s wealth. Follow https://pmc.ncbi.nlm.nih.gov/articles/PMC5348557/ and https://rpubs.com/Sternonyos/526030

Code 
library(psych)

df$ownership <- ifelse(df$OWNERSHIP == 1,1,0)
df$watsup <- ifelse(df$WATSUP == 10,1,0)
df$aircon <- ifelse(df$AIRCON == 10,0,1)
df$computer <- ifelse(df$COMPUTER == 2,1,0)
df$washer <- ifelse(df$WASHER == 2,1,0)
df$refrig <- ifelse(df$REFRIG == 2,1,0)
df$tv <- ifelse(df$TV == 20,1,0)
df$radio <- ifelse(df$RADIO == 2,1,0)
df$electric <- ifelse(df$ELECTRIC == 1,1,0)
df$hotwater <- ifelse(df$HOTWATER == 2,1,0)
df$phone <- ifelse(df$PHONE == 2,1,0)

wealth_index <- df[,c("qh","ownership","watsup","aircon","computer",
                      "washer","refrig","tv","radio","electric",
                      "hotwater","phone")] 

prn<- principal(wealth_index[,-1], rotate="varimax", 
                      nfactors=3,covar=T, scores=TRUE)
index=prn$scores[,1]
nlab<-c(1,2,3,4,5)
newdata<-mutate(wealth_index,
                quintile=as.factor(cut(index,breaks=5,labels=nlab)))

wealdf <- newdata %>% group_by(qh,quintile) %>% count()

wealdf$quintile <-  case_when(
  wealdf$quintile == 1 ~ "Poorest, 20%",
  wealdf$quintile == 2 ~ "Near poor, 20%",
  wealdf$quintile == 3 ~ "Middle, 20%",
  wealdf$quintile == 4 ~ "Richer, 20%",
  wealdf$quintile == 5 ~ "Richest, 20%") %>% 
  factor(levels = c("Poorest, 20%", 
                    "Near poor, 20%",
                    "Middle, 20%",
                    "Richer, 20%",
                    "Richest, 20%")
  )

wealdf %>% scale_per() %>% 
  ggplot() +
  geom_col(aes(x = per,
               y = quintile)) +
  facet_wrap(vars(qh),
             # scales = "free",
             ncol = 5)+
  labs(x = "Percentage of total population(%)",
       y = "Socioeconomic status (quintile)")+
  theme_light()+
  theme(axis.text.x = element_text(size = 6),
        axis.text.y = element_text(size = 6))