Project4_CarolineAllan.R

## Caroline Allan
## U.S. History Through Census Data
## Project 4
## Children with Divorced Parents 1940-2000
## March 1, 2016

rm(list = ls())

# load packages
library(readr)
library(dplyr)
library(ggplot2)
library(scales)
library(ggthemes)
library(tidyr)

## Set Working Directory
setwd("~/Desktop/HistThroughCensusData/Project4")

## Read In Data
data_raw <- read_csv('usa_00016.csv')

## Remove Alaska and Hawaii before 1960, GQ = 1
a <- data_raw %>% 
  filter(STATEFIP < 60 & (YEAR >= 1960 | !(STATEFIP %in% c(2, 15)))) %>% 
  filter(GQ == 1)

## Create dataset of children <18
children <- a %>% 
  filter(AGE < 18) %>% 
  select(YEAR, SERIAL, PERNUM, PERWT, MOMLOC, POPLOC, AGE, RACE, HISPAN)

## Parent dataset
parent <- a %>% 
  select(YEAR, SERIAL, PERNUM, MARST)

## Dataset with mom's marital status (if kid lives with mom)
mom <- merge(children, parent, by.x = c('YEAR', 'SERIAL', 'MOMLOC'), 
             by.y = c('YEAR', 'SERIAL', 'PERNUM'))
mom <- mom %>% 
  rename(MOM_MARST = MARST)%>% 
  select(YEAR, SERIAL, PERNUM, MOM_MARST)
head(mom)

## Dataset with dad's marital status (if kid lives with dad)
dad <- merge(children, parent, by.x = c('YEAR', 'SERIAL', 'POPLOC'), 
             by.y = c('YEAR', 'SERIAL', 'PERNUM'))
dad <- dad %>% 
  rename(DAD_MARST = MARST) %>% 
  select(YEAR, SERIAL, PERNUM, DAD_MARST)
head(dad)

## Merge mom marital status in with child data
data_final1 <- merge(children, mom, by = c('YEAR', 'SERIAL', 'PERNUM'), all.x = TRUE)

## Merge dad marital status in with child data
data_final2 <- merge(data_final1, dad, by = c('YEAR', 'SERIAL', 'PERNUM'), all.x = TRUE)

## Add in divorced variable
data_final3 <- data_final2 %>% 
  mutate(divorced = ifelse(data_final2$MOM_MARST == 4 | data_final2$DAD_MARST == 4, 
                           "divorced", "not divorced"))

## Select necessary variables, label race categories
data_final4 <- data_final3 %>% 
  select(YEAR, PERWT, AGE, RACE, HISPAN, divorced) %>% 
  mutate(RACE = ifelse(HISPAN == 1, "Non-White",
                       ifelse(RACE == 1, "White",
                              "Non-White")))

## Create divorced and not divorced variables
data_final5 <- data_final4 %>% 
  mutate(divorce = ifelse(divorced == 'divorced', PERWT, 0)) %>% 
  mutate(notdivorce = ifelse(divorced == 'not divorced', PERWT, 0)) %>% 
  filter(!is.na(divorced))

## Save data to save future compile time
write.csv(data_final5, "finaldata.csv")

## Plot 1 and 2 Data
byyear <- data_final5 %>% 
  group_by(YEAR) %>% 
  summarize(NumDivorced = sum(divorce),
            NumNot = sum(notdivorce)) %>% 
  mutate(Percent = NumDivorced/(NumDivorced + NumNot))

## Plot 3 and 4 Data
byyearraceage <- data_final5 %>%
  mutate(AGECAT = factor(cut(AGE, c(0, 6, 12, 18), right = FALSE))) %>% 
  group_by(YEAR, RACE, AGECAT) %>% 
  summarize(NumDivorced = sum(divorce),
            NumNot = sum(notdivorce)) %>% 
  mutate(Percent = NumDivorced/(NumDivorced + NumNot))

## Plot 1
## Create a plot of # kids with divorced parents by year
plot1 <- ggplot(byyear, aes(x = YEAR, y = NumDivorced)) +
  geom_bar(stat = 'identity', fill = "#33CCCC", color = "black") +
  theme_few() +
  labs(x = "Year",
       y = "Number of Children Living With a Divorced Parent",
       title = "Estimated Number of Children 0-17 Living with a\nDivorced Parent 1940-2000") +
  scale_x_continuous(breaks = c(1940, 1950, 1960, 1970, 1980, 1990, 2000)) +
  scale_y_continuous(labels = comma)
print(plot1)
ggsave(file="NumKids.png",dpi=300)


## Plot 2
## Create a plot of % kids with divorced parents by year
plot2 <- ggplot(byyear, aes(x = YEAR, y = Percent)) +
  geom_bar(stat = 'identity', fill = "#33CCCC", color = "black") +
  theme_few() +
  labs(x = "Year",
       y = "Percent of Children Living With a Divorced Parent",
       title = "Estimated Percent of Children 0-17 Living with a\nDivorced Parent 1940-2000") +
  scale_x_continuous(breaks = c(1940, 1950, 1960, 1970, 1980, 1990, 2000)) +
  scale_y_continuous(labels = percent)
print(plot2)
ggsave(file="PercKids.png",dpi=300)


## Plot 3
## Create a plot of # kids with divorced parents by year, race, and age
plot3 <- ggplot(byyearraceage, aes(x = YEAR, y = NumDivorced, colour = AGECAT)) +
  geom_line() +
  theme_few() +
  facet_grid(RACE~.) +
  scale_color_brewer(name = 'Age\nCategory',
                     labels = c('0-5', '6-11', '11-17'),
                     palette = "Set2",
                     guide = guide_legend(reverse=TRUE)) +
  labs(x = "Year",
       y = "Number of Children Living With a Divorced Parent",
       title = "Estimated Number of Children 0-17 Living with a\nDivorced Parent 1940-2000 by Age and Race") +
  scale_x_continuous(breaks = c(1940, 1950, 1960, 1970, 1980, 1990, 2000)) +
  scale_y_continuous(labels = comma)
print(plot3)
ggsave(file="NumKidsRaceAge.png",dpi=300)


## Plot 4
## Create a plot of % kids with divorced parents by year, race, and age
plot4 <- ggplot(byyearraceage, aes(x = YEAR, y = Percent, colour = AGECAT)) +
  geom_line() +
  theme_few() +
  scale_color_brewer(name = 'Age\nCategory',
                    labels = c('0-5', '6-11', '11-17'),
                    palette = "Set2",
                    guide = guide_legend(reverse=TRUE)) +
  facet_grid(RACE~.) +
  labs(x = "Year",
       y = "Percent of Children Living With a Divorced Parent",
       title = "Estimated Percent of Children 0-17 Living with a\nDivorced Parent 1940-2000 by Age and Race") +
  scale_x_continuous(breaks = c(1940, 1950, 1960, 1970, 1980, 1990, 2000)) +
  scale_y_continuous(labels = percent)
print(plot4)
ggsave(file="PercKidsRaceAge.png",dpi=300)