Vote count plots

library(readr)
library(dplyr)
library(ggplot2)
library(ggrepel)

download.file('https://raw.githubusercontent.com/alex/nyt-2020-election-scraper/master/battleground-state-changes.csv', destfile = 'battleground-state-changes.csv')
dat <- read_csv('battleground-state-changes.csv')

dat <- dat %>%
  group_by(state) %>%
  arrange(timestamp) %>%
  mutate(
    state = gsub(" .*", "", state),
    biden_votes = ifelse(leading_candidate_name == "Biden", leading_candidate_votes, trailing_candidate_votes),
    trump_votes = ifelse(leading_candidate_name == "Trump", leading_candidate_votes, trailing_candidate_votes),
    total_counted = biden_votes + trump_votes,
    total = last(total_counted) + last(votes_remaining),
    percent_biden = biden_votes / total_counted * 100,
    percent_trump = trump_votes / total_counted * 100,
    percent_counted = total_counted / total * 100,
    advantage_trump = percent_trump - percent_biden
  )



state <- dat %>%
  filter(state == "Arizona")
ggplot(state, aes(x = percent_counted, y = advantage_trump)) +
  ggtitle("Arizona") +
  geom_point() +
  stat_smooth(geom = "line", method = lm, fullrange = TRUE, se = FALSE,
    alpha = 0.2, color ="blue") +
  expand_limits(y = 0, x = 100) +
  ylab("Advantage Trump\n(positive = Trump is winning)") +
  xlab("Percent of ballots counted") +
  geom_hline(yintercept = 0, linetype = "dashed") +
  geom_vline(xintercept = 100) +
  geom_label_repel(
    aes(label = strftime(timestamp, "%m-%d %H:%M ", tz ="EST", usetz = TRUE)),
    data = state %>%
      filter(timestamp == min(timestamp) | timestamp == max(timestamp)),
    segment.size  = 0.2,
    segment.color = "grey50",
    color         = "grey60",
    hjust         = c(-0.5, 1.5)
  ) +
  scale_y_continuous(breaks = scales::breaks_extended(15, Q = c(1, 2, 5))) +
  scale_x_continuous(breaks = scales::breaks_extended(12, Q = c(1, 2, 5))) +
  theme_bw()

state <- dat %>%
  filter(state == "Georgia")
ggplot(state, aes(x = percent_counted, y = advantage_trump)) +
  ggtitle("Georgia") +
  geom_point(size = 1) +
  stat_smooth(geom = "line", method = lm, fullrange = TRUE, se = FALSE,
    alpha = 0.2, color ="blue") +
  expand_limits(y = 0, x = 100) +
  ylab("Percent advantage Trump") +
  xlab("Percent of ballots counted") +
  geom_hline(yintercept = 0, linetype = "dashed") +
  geom_vline(xintercept = 100) +
  geom_label_repel(
    aes(label = strftime(timestamp, "%m-%d %H:%M ", tz ="EST", usetz = TRUE)),
    data = state %>%
      filter(timestamp == min(timestamp) | timestamp == max(timestamp)),
    segment.size  = 0.2,
    segment.color = "grey50",
    color         = "grey60",
    hjust         = c(-0.5, 1.5)
  ) +
  scale_y_continuous(breaks = scales::breaks_extended(15, Q = c(1, 2, 5))) +
  scale_x_continuous(breaks = scales::breaks_extended(12, Q = c(1, 2, 5))) +
  theme_bw()

state <- dat %>%
  filter(state == "Pennsylvania")
ggplot(state, aes(x = percent_counted, y = advantage_trump)) +
  ggtitle("Pennsylvania") +
  geom_point(size = 1) +
  stat_smooth(geom = "line", method = lm, fullrange = TRUE, se = FALSE,
    alpha = 0.2, color ="blue") +
  expand_limits(y = -3, x = 100) +
  ylab("Percent advantage Trump") +
  xlab("Percent of ballots counted") +
  geom_hline(yintercept = 0, linetype = "dashed") +
  geom_vline(xintercept = 100) +
  geom_label_repel(
    aes(label = strftime(timestamp, "%m-%d %H:%M ", tz ="EST", usetz = TRUE)),
    data = state %>%
      filter(timestamp == min(timestamp) | timestamp == max(timestamp)),
    segment.size  = 0.2,
    segment.color = "grey50",
    color         = "grey60",
    hjust         = c(-0.5, 1.5)
  ) +
  scale_y_continuous(breaks = scales::breaks_extended(15, Q = c(1, 2, 5))) +
  scale_x_continuous(breaks = scales::breaks_extended(12, Q = c(1, 2, 5))) +
  theme_bw()

state <- dat %>%
  filter(state == "Nevada")
ggplot(state, aes(x = percent_counted, y = advantage_trump)) +
  ggtitle("Nevada") +
  geom_point() +
  stat_smooth(geom = "line", method = lm, fullrange = TRUE, se = FALSE,
    alpha = 0.2, color ="blue") +
  expand_limits(y = 0, x = 100) +
  ylab("Percent advantage Trump") +
  xlab("Percent of ballots counted") +
  geom_hline(yintercept = 0, linetype = "dashed") +
  geom_vline(xintercept = 100) +
  geom_label_repel(
    aes(label = strftime(timestamp, "%m-%d %H:%M ", tz ="EST", usetz = TRUE)),
    data = state %>%
      filter(timestamp == min(timestamp) | timestamp == max(timestamp)),
    segment.size  = 0.2,
    segment.color = "grey50",
    color         = "grey60",
    hjust         = c(1.5, -0.5)
  ) +
  scale_y_continuous(breaks = scales::breaks_extended(15, Q = c(1, 2, 5))) +
  scale_x_continuous(breaks = scales::breaks_extended(12, Q = c(1, 2, 5))) +
  theme_bw()

state <- dat %>%
  filter(state == "Alaska")
ggplot(state, aes(x = percent_counted, y = advantage_trump)) +
  ggtitle("Alaska") +
  geom_point() +
  stat_smooth(geom = "line", method = lm, fullrange = TRUE, se = FALSE,
    alpha = 0.2, color ="blue") +
  expand_limits(y = 0, x = 100) +
  ylab("Percent advantage Trump") +
  xlab("Percent of ballots counted") +
  geom_hline(yintercept = 0, linetype = "dashed") +
  geom_vline(xintercept = 100) +
  geom_label_repel(
    aes(label = strftime(timestamp, "%m-%d %H:%M ", tz ="EST", usetz = TRUE)),
    data = state %>%
      filter(timestamp == min(timestamp) | timestamp == max(timestamp)),
    segment.size  = 0.2,
    segment.color = "grey50",
    color         = "grey60",
    hjust         = c(1.5, -0.5)
  ) +
  scale_y_continuous(breaks = scales::breaks_extended(15, Q = c(1, 2, 5))) +
  scale_x_continuous(breaks = scales::breaks_extended(12, Q = c(1, 2, 5))) +
  theme_bw()