electionViz

R package for visualizations of election (or poll) results as easy as adding geom_electoral_building. While the geom is still in the planning stages, several types of visualizations are implemented and ready to use.

Installation

The development version from GitHub with:

# install.packages("devtools")
devtools::install_github("heike/electionViz")

Example

We will be inundated with information about the state of the polls in the build-up to the upcoming Presidential election of 2020.

With the tools of this package you will be able to pick your favorite visualization(s) and explore the results your own way.

# load the package
library(electionViz)
library(tidyverse)

Hexbin Cartogram of Election Results by State

Each state is represented by one hexagon. This map has been made available by Andrew X Hill at CARTO.

data(elections)
el12 <- elections %>% filter(year == 2012)
gg12 <- hexplot(el12$state, el12$perc_rep > el12$perc_dem) +
  scale_fill_party("", labels=c("Democrat", "Republican")) +
  theme(legend.position = "bottom") +
  ggtitle("Election 2012")

el16 <- elections %>% filter(year == 2016)
gg16 <- hexplot(el16$state, el16$perc_rep > el16$perc_dem) +
  scale_fill_party("", labels=c("Democrat", "Republican")) +
  theme(legend.position = "bottom") +
  ggtitle("Election 2016")

gridExtra::grid.arrange(gg12, gg16, ncol=2)

Hexbin Cartogram of the US Presidential Election by Electoral Votes

Each state is represented by a set of hexagons corresponding in number to the state’s electoral votes. This map was adapted from the object sf_FiveThirtyEightElectoralCollege in Bhaskar Karambelkar’s R package tilegramsR.

Electoral Building of the U.S. Presidential Election by Electoral Votes and Margin of Victory

“Electoral Building” diagram, inspired by a 2000 graphic from the New York Times. Vote margin is represented on the x axis, while number of electoral votes is represented in the height.

Visualization of Polls

Accessing data of the most recent polls from RealClearPolitics or FiveThirtyEight is done with functions rcp_update() or fivethirtyeight_update() respectively. Each one of these functions has a parameter polls to allow for a finer grained choice of which polls to focus on:

new_polls <- fivethirtyeight_update(polls="president_polls")
#> Warning: 44 parsing failures.
#>  row               col           expected          actual                                                                  file
#> 1923 sponsor_candidate 1/0/T/F/TRUE/FALSE Nicole Galloway 'https://projects.fivethirtyeight.com/polls-page/president_polls.csv'
#> 1924 sponsor_candidate 1/0/T/F/TRUE/FALSE Nicole Galloway 'https://projects.fivethirtyeight.com/polls-page/president_polls.csv'
#> 2585 sponsor_candidate 1/0/T/F/TRUE/FALSE Al Gross        'https://projects.fivethirtyeight.com/polls-page/president_polls.csv'
#> 2586 sponsor_candidate 1/0/T/F/TRUE/FALSE Al Gross        'https://projects.fivethirtyeight.com/polls-page/president_polls.csv'
#> 2968 sponsor_candidate 1/0/T/F/TRUE/FALSE Raphael Warnock 'https://projects.fivethirtyeight.com/polls-page/president_polls.csv'
#> .... ................. .................. ............... .....................................................................
#> See problems(...) for more details.
new_polls %>% filter(!is.na(state)) %>%
  select(poll_id, start_date, end_date, pollster, state, candidate_name, pct)
#> # A tibble: 6,870 x 7
#>    poll_id start_date end_date   pollster    state       candidate_name      pct
#>      <dbl> <date>     <date>     <chr>       <chr>       <chr>             <dbl>
#>  1   71295 2020-10-13 2020-10-16 YouGov      Arizona     Joseph R. Biden …    50
#>  2   71295 2020-10-13 2020-10-16 YouGov      Arizona     Donald Trump         47
#>  3   71296 2020-10-13 2020-10-16 YouGov      Wisconsin   Joseph R. Biden …    51
#>  4   71296 2020-10-13 2020-10-16 YouGov      Wisconsin   Donald Trump         46
#>  5   71242 2020-09-19 2020-10-16 SurveyMonk… Wyoming     Joseph R. Biden …    33
#>  6   71242 2020-09-19 2020-10-16 SurveyMonk… Wyoming     Donald Trump         66
#>  7   71242 2020-09-19 2020-10-16 SurveyMonk… Wyoming     Joseph R. Biden …    33
#>  8   71242 2020-09-19 2020-10-16 SurveyMonk… Wyoming     Donald Trump         66
#>  9   71243 2020-09-19 2020-10-16 SurveyMonk… West Virgi… Joseph R. Biden …    35
#> 10   71243 2020-09-19 2020-10-16 SurveyMonk… West Virgi… Donald Trump         63
#> # … with 6,860 more rows

Polls by State

The difference in percentage between democratic and republican percentage of the last five polls (of likely or registered voters) for each state are shown as grey points, colored points show average difference for each point. States are ordered according to difference in percentage. The yellow rectangle shows a margin of +/- 5 percent - the typical margin of error of a poll.

Building of polls

state_3 <- state_3 %>% filter(!(state %in% c("Maine", "Nebraska")))
electoral_building(
    state_district = state_3$state, 
    electoral_votes = state_3$electoral_votes, 
    perc_dem = state_3$perc_dem, 
    perc_rep = state_3$perc_rep,
    source = state_3$source) +
  scale_color_party("Party") +
  scale_fill_party("Party") +
  theme(legend.position = "none") +
  ggtitle("Electoral building, based on polls")

A snake of beads

bead_snake_plot(
   electoral_votes_2016$state_district,
   electoral_votes_2016$electoral_votes,
   electoral_votes_2016$perc_dem,
   electoral_votes_2016$perc_rep, 
  height = 30, buffer = 3) +
  ggtitle("Election 2016")
#> Joining, by = "state_district"