2020 Presidential Election Returns from Decision Desk

Overview

Decision Desk HQ tweeted out election returns for the 2020 presidential election on the night of the election and the days that followed from their Twittter account.

This repo preserves those tweets and wrangles the information into returns by time, state, and candidate.

Details

01-get-tweets.R downloads and saves Decision Desk HQ’s Twitter timeline, which contains important tweets with the vote counts at particular point in time for important states. This scripts writes the raw timeline as raw/decision-desk-timeline.rds and a slightly altered version as raw/decision-desk-timeline.csv.
02-wrangle-tweets.R takes the raw tweets from protected/ (not raw/, see note below) and extracts the relevant information. It saves the data as returns.rds and returns.csv. Because of the dttm objects, RDS works a bit better here.

The scripts do not automatically write to the protected/ directory. I create this directory by manually copying the raw/ directory. This protects against accidentally deleting the raw tweets that contain the relevant information.

library(tidyverse)
library(lubridate)

returns <- read_rds("returns.rds") %>%
  glimpse()

## Rows: 468
## Columns: 5
## $ state     <chr> "AZ", "AZ", "GA", "GA", "GA", "GA", "AZ", "AZ", "AZ", "AZ",…
## $ time      <dttm> 2020-11-10 02:09:35, 2020-11-10 02:09:35, 2020-11-10 02:08…
## $ candidate <chr> "Biden", "Trump", "Biden", "Trump", "Biden", "Trump", "Bide…
## $ votes     <dbl> 1648642, 1633896, 2469118, 2456781, 2467748, 2456157, 16434…
## $ text      <chr> "AZ Presidential Election Results\n\nBiden (D): 49.47% (1,6…

Read Directly from Web

You can read the RDS data directly from GitHub with rio::import().

r <- rio::import("https://github.com/carlislerainey/decision-desk-hq-returns/raw/master/returns.rds")

GA Returns Example

ga <- returns %>%
  filter(state == "GA") %>%
  filter(time > ymd("2020-11-05"))

ggplot(ga, aes(x = time, y = votes, color = candidate)) + 
  geom_line() + 
  scale_y_log10() + 
  scale_color_manual(values = c("Biden" = scales::muted("blue"),
                                "Trump" = scales::muted("red"))) + 
  labs(title = "Presidental Election Returns in GA") + 
  theme_bw()