library(haven)
library(dplyr)
library(ggplot2)

raw_welfare<-read_dta(file="D:/R/Koweps/koweps_h15_2020_beta1.dta", encoding = "latin1")
welfare<-raw_welfare
#str(welfare)
welfare<-rename(welfare, gender=h1501_4, birth=h1501_5, edu=h1501_6, marriage=h1501_11, income=h15_cin, code_region=h15_reg7)
welfare<-mutate(welfare, salary=(income/12))
welfare$salary
class(welfare$gender)
table(welfare$gender)

# 성별에 따른 월급 차이
welfare$gender<-ifelse(welfare$gender==1, "male","female")
table(welfare$gender)
qplot(welfare$gender)
summary(welfare$salary)
qplot(welfare$salary, bins=30)
qplot(welfare$salary)+xlim(0,1000)
gender_salary<-welfare %>% group_by(gender) %>% summarise(mean_salary=mean(salary))
gender_salary
ggplot(data=gender_salary, aes(x=gender, y=mean_salary))+geom_col()

# 나이와 월급의 관계
summary(welfare$birth)
qplot(welfare$birth)
welfare$age<-2020-welfare$birth+1 # 나이
summary(welfare$age)
qplot(welfare$age)
age_salary<-welfare %>% group_by(age) %>% summarise(mean_salary=mean(salary))
ggplot(data=age_salary, aes(x=age, y=mean_salary))+geom_line() # 나이에 따른 월급 평균

# 연령대와 월급의 관계
welfare$ageg<-ifelse(welfare$age < 30, "청소년", ifelse(welfare$age <= 60, "중장년", "노년"))
welfare$age
welfare$ageg
ageg_salary<-welfare %>% group_by(ageg) %>% summarise(mean_salary=mean(salary))
ageg_salary
ggplot(data=ageg_salary, aes(x=ageg, y=mean_salary))+geom_col()

# 성별 및 연령대 월급의 관계
gender_ageg_salary<-welfare %>% group_by(ageg,gender) %>%  summarise(mean_salary=mean(salary),.groups='drop') 
gender_ageg_salary
ggplot(data=gender_ageg_salary, aes(x=ageg, y=mean_salary, fill=gender))+geom_col()
ggplot(data=gender_ageg_salary, aes(x=ageg, y=mean_salary, fill=gender))+geom_col(position="dodge")

# 성별 및 연령별 월급 평균 
gender_age<-welfare %>% group_by(age, gender) %>% summarise(mean_salary=mean(salary),.groups='drop')
gender_age
ggplot(data=gender_age, aes(x=age, y=mean_salary, col=gender))+geom_line()

# 교육수준 및 월급의 관계 
summary(welfare$edu)
welfare$edug<-ifelse(welfare$edu >= 8, "대학원", ifelse(welfare$edu >= 6, "대학", 
                                                      ifelse(welfare$edu >= 4, "중등", "초등")))
edug_salary<-welfare %>% group_by(edug) %>% summarise(mean_salary=mean(salary))
edug_salary
ggplot(data=edug_salary, aes(x=edug, y=mean_salary))+geom_col()

# 성별 및 교육수준별 월급 평균
gender_edug_salary<-welfare %>% group_by(edug,gender) %>%  summarise(mean_salary=mean(salary),.groups='drop') 
gender_edug_salary
ggplot(data=gender_edug_salary, aes(x=edug, y=mean_salary, fill=gender))+geom_col()+
  scale_x_discrete(limit=c("대학","대학원","중등","초등"))
ggplot(data=gender_edug_salary, aes(x=edug, y=mean_salary, fill=gender))+geom_col(position="dodge")+
  scale_x_discrete(limit=c("대학","대학원","중등","초등"))

# 지역별 연령대 비율
class(welfare$code_region)
table(welfare$code_region)
list_region <- data.frame(code_region = c(1:7),  region= c("서울",
                                                          "수도권(인천.경기)",
                                                          "부산/경남/울산",
                                                          "대구/경북",
                                                          "대전/충남",
                                                          "강원/충북",
                                                          "광주/전남/전북/제주"))
list_region
welfare <- left_join(welfare, list_region, id="code_region")
welfare %>% select(code_region, region)

region_ageg_1 <- welfare %>% group_by(region,ageg) %>%  
                           summarise(n=n(),.groups='drop') %>%
                           mutate(tot_group = sum(n)) %>%
                           mutate(pct=round(n/tot_group*100, 2))
head(region_ageg_1)                                   
ggplot(data=region_ageg_1, aes(x=region, y=pct, fill=ageg))+geom_col()+coord_flip()

region_ageg_2 <- welfare %>% count(region, ageg) %>%  
                             group_by(region) %>%
                             mutate(pct=round(n/sum(n)*100, 2))
head(region_ageg_2)                                   
ggplot(data=region_ageg_2, aes(x=region, y=pct, fill=ageg))+geom_col()+coord_flip()

# 노년층 비율 높은 순으로 정렬하기
list_order_old_1 <- region_ageg_1 %>% 
                    filter(ageg =="노년") %>%
                    arrange(pct)
list_order_old_1
ggplot(data=list_order_old_1, aes(x=reorder(region,pct), y=pct, fill=ageg))+geom_col()+coord_flip()

list_order_old_2 <- region_ageg_2 %>% 
                    filter(ageg =="노년") %>%
                    arrange(pct)
list_order_old_2
ggplot(data=list_order_old_2, aes(x=region, y=pct, fill=ageg))+geom_col()+coord_flip()
ggplot(data=list_order_old_2, aes(x=reorder(region,pct), y=pct, fill=ageg))+geom_col()+coord_flip()


# 청소년층 비율 높은 순으로 정렬하기
list_order_young_1 <- region_ageg_1 %>% 
                    filter(ageg =="청소년") %>%
                    arrange(pct)
list_order_young_1
ggplot(data=list_order_young_1, aes(x=region, y=pct, fill=ageg))+geom_col()+coord_flip()
ggplot(data=list_order_young_1, aes(x=reorder(region,pct), y=pct, fill=ageg))+geom_col()+coord_flip()

list_order_young_2 <- region_ageg_2 %>% 
                    filter(ageg =="청소년") %>%
                    arrange(pct)
list_order_young_2
ggplot(data=list_order_young_2, aes(x=region, y=pct, fill=ageg))+geom_col()+coord_flip()
ggplot(data=list_order_young_2, aes(x=reorder(region,pct), y=pct, fill=ageg))+geom_col()+coord_flip()