# Sample Reading Data -- ACS 
# This is an example for only illustrate to use the ACS without considering the weights on the individual personal record
# Heavy modifications are necessary to use as an actual research
# Ideas from Tom Bales (2021 Winter Class)

rm(list = ls())

acs_pop_il <- read_csv("/var/www/html/jlee141/econdata/acs_data/acs2019_pop_il.csv")
str(acs_pop_il)
library(dplyr)

# Select only variables you need here to be used for the project.
# Find the variables you are interested in from see PUMS_DATA_Dictionary_2018.pdf after page 30
# Here is an example to include 8 variables and only BA or higher education attainment:

acs_pop_il2 <- acs_pop_il %>%  
  select(ST,SCHL,AGEP,CIT,FOD1P,WAGP,SEX,OCCP) %>% 
  mutate(SEX = as.factor(SEX), 
         CIT = as.factor(CIT), 
         SCHL = as.factor(SCHL))

# Filter out the Non-college / Graduate school degrees attained #
# Select only people has Associate and Bachelor Degree - see PUMS_DATA_Dictionary_2018.pdf Page 42
acs_pop_il2 <- acs_pop_il2 %>% filter(SCHL %in% c("21","22","23","24")) 

# Degrees from FOD1P. Check the full list in the dictionary. Dummy for Business Major

acs_pop_il2 <- acs_pop_il2 %>%  
  mutate(Degree= case_when(
    FOD1P %in% c(6200,6201,6202,6203,6204,6205,6206,6207,6209,6210,6211,6212,6299) ~ "BUSINESS" ,  
    FOD1P %in% c(2100,2101,2102,2105,2106,2107,3700,3701,3702,4005) ~ "COMP_MATH",  
    TRUE ~ "A.OTHERS")) %>%   
  mutate(Degree = as.factor(Degree), 
         OLDER  = ifelse(acs_pop_il2$AGEP >55,1,0),
         AGE2 = AGEP^2,
         logWage = log(WAGP)) %>% 
  filter(WAGP > 0)
  
# Let's find if the degree matter to the wage income
lm1 <- lm(logWage~SEX+SCHL+AGEP+AGE2+CIT+Degree, data=acs_pop_il2)
summary(lm1)