elections-chart-2016.R

library(tidyverse)
library(ggrepel)
library(scales)
results <- read_csv("results.csv")

# add GDP rank, winning candidate, chart labels
results <- results %>% 
  arrange(desc(GDP)) %>%
  mutate(GDPRank = 1:51) %>%
  mutate(Win = ifelse(ClintonNum>TrumpNum, "Clinton", "Trump")) %>%
  mutate(Label = paste0(State_Abb, " [", `Electoral Votes`, "]"))
  
# create tidy data, separate camelcase column
results_tidy <- results %>% 
  gather(key, value, -State, -`Electoral Votes`, -GDP, -State_Abb, -GDPRank, -Win, -Label) %>%
  separate(key, c("Candidate", "VoteType"), "(?<=.)(?=[A-Z])") %>%
  spread(VoteType, value)

# scatter plot
ggplot(results_tidy %>% filter(Candidate=="Clinton")) + 
  geom_point(aes(x=GDP, y=Pct, size=`Electoral Votes`, col=Win, alpha=0.5)) + 
  scale_color_manual(values=c("#0052A5", "#E0162B")) +
  geom_hline(yintercept=0.5, col="#404040", linetype=2) +
  annotate("text", x=2.5e6, y=0.515, label="50% Line", col="#404040") +
  labs(title="2016 Election Results vs. GDP by State") + 
  scale_y_continuous(name="Percent of Votes for Clinton", labels=percent) +
  scale_x_log10(name="GDP (in millions $USD) from Wikipedia",
                breaks=trans_breaks("log10", function(x) 10^x),
                labels=trans_format("log10", math_format(10^.x))) +
  geom_text_repel(
                  # data=subset(results_tidy %>% filter(Candidate=="Clinton"), GDPRank<=10), 
                  aes(x=GDP, y=Pct, label=Label),
                  fontface="bold") + theme_bw(base_size=12) + 
  theme(legend.position="none") + 
  annotate("rect", xmin=0.9e6, xmax=3e6, ymin=0.865, ymax=0.93, fill="#FFFBCC") +
  annotate("text", x=1e6, y=0.90, 
           label="Labels: State [Electoral Votes]\nSize: Number of Electoral Votes", 
           col="#404040", fontface="bold.italic", hjust=0)