Exploring US Demographics with data from the General Social Survey

Graphing Income by Race

GSS <- GSS[ which(!is.na(GSS$race) & !is.na(GSS$coninc)), ] #Removing NA values

ggplot(data =GSS, aes(x=race, y=coninc)) + #Choosing variables race and family income
  geom_boxplot() +
  ggtitle('Race and Income Boxplot') + #Title of the plot
  ylab("Family Income $") + xlab("Race")+ #Labels for the axis
  theme(plot.title = element_text(hjust = 0.5)) #Centering the title

This boxplot displays the distribution of income by race. This has profuond insights, such as the fact that Black people have the lowest mean income, with “other” having the middle average, and White people having the highest income.

Graphing Race, Age, and TV Hours

gss_cat %>%
  group_by(race, age) %>%
  summarise(meantv=mean(tvhours, na.rm = T)) %>% 
  #Using dplyr to create a mean TV hours variable and summarise it by race and age
  ggplot(aes(x=age, y=meantv,color=race))+ #GGplot x-axis is age, y-axis is Mean TV Hours, and Color is race
  geom_point() + #Speceifying a scatterplot
  geom_smooth(method = "lm")+ #Affs a trend line on the plot
  ylab("Hours of TV per Week") + xlab("Age")+ #Labels for the axis
  ggtitle("Age, Hours of TV, and Race Scatterplot")+ #Title of the plot
  theme(plot.title = element_text(hjust = 0.5)) #Centering the title

The graph shows that for White, Black, and other races the hours of TV watched per week rises as that population gets older. The three seperate trend lines also reveal racial differences. Whites and others watch similiar amounts of TV especially as they get older, Black people watched signifcantly more TV on average than the other races.

Graphing Race and Political Party

GSS$partyid<-ifelse(GSS$partyid == "Strong republican", "Republican",ifelse(GSS$partyid == "Not str republican", "Republican", ifelse(GSS$partyid == "Ind, near rep", "Republican", ifelse(GSS$partyid == "Ind, near dem", "Democrat", ifelse(GSS$partyid == "Not str democrat", "Democrat",ifelse(GSS$partyid == "Strong Democrat", "Democrat", "Independent" )))))) #Using an ifelse ladder to combine political party identification into Democrat, Repubican, Independent


ggplot(GSS, aes(partyid, fill = race)) + #Selecting political party as variable and race as group value
  theme(axis.text.x = element_text(angle = 45, hjust = 1))+  #This line tilts the axis names so they are readable
  geom_bar(position = "dodge")+ #This splits the bar chart by race instead of stacking the bars
  geom_text(aes(label=..count..),stat="count",position=position_dodge(width=0.9), vjust=-0.25)+ #Adds labels to each bar
   ylab("Count") + xlab("Political Party")+ #Labels for the axis
  ggtitle("Race and Political Party Bar Chart")+ #Title of the plot
  theme(plot.title = element_text(hjust = 0.5)) #Centering the title

This graph reveals that Democrats are the most racially diverse political group, while Independent is the most popular political category.

Income

mean_inc <- aggregate(coninc~race, data=GSS, mean) #Use aggregate function to take mean of family income for each race
print(mean_inc)

##    race   coninc
## 1 Black 28445.62
## 2 Other 38010.33
## 3 White 47727.90

This information shows that the mean income for the White race is 1.7 times more than the mean income for the Black race, clearly there is a racial income disparity in the United States.

TV Hours

One of the most evident correlations was between TV Hours and Age. This relationship can further be explored by running a regression on the two variables.

tvhours_reg<-lm(tvhours~age + coninc, data=GSS) #Running a linear regression on TV hours and Age
print(tvhours_reg) #Pritning results of the regression

## 
## Call:
## lm(formula = tvhours ~ age + coninc, data = GSS)
## 
## Coefficients:
## (Intercept)          age       coninc  
##   1.647e+00    4.067e-02   -1.371e-05

summary(tvhours_reg) #Summarising the regression

## 
## Call:
## lm(formula = tvhours ~ age + coninc, data = GSS)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.9414 -1.4945 -0.5185  0.8314 20.3600 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.647e+00  1.929e-01   8.541   <2e-16 ***
## age          4.067e-02  3.485e-03  11.669   <2e-16 ***
## coninc      -1.371e-05  1.480e-06  -9.267   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.665 on 1858 degrees of freedom
##   (1006 observations deleted due to missingness)
## Multiple R-squared:  0.1079, Adjusted R-squared:  0.1069 
## F-statistic: 112.3 on 2 and 1858 DF,  p-value: < 2.2e-16

mean_tv <- aggregate(tvhours~race, data=GSS, mean) #Use Aggregate to find mean TV Hours per Race
print(mean_tv)

##    race  tvhours
## 1 Black 4.046875
## 2 Other 2.697297
## 3 White 2.840348

The regression allows us to see the relationship between age and hours of TV one watches. The regression shows that an increase in age by 1 year results in a 0.04 increase in the hours of TV someone watches. An increase in income by 1,000 is associated with 0.015 decrease in the amount of TV one watches. However; the R-squared value shows that these variables only explain 10% of the data observations. From this data we can also compare the mean hours of TV each race watches.