#==============================================================================#

# Library Load-in---------------------------------------------------------------

#==============================================================================#

library(tidyverse) # Data Wrangling and Plotting

library(here) # Files location and loading

library(summarytools) # Exploratory Data Analysis

library(colorfindr) # To get colour palettes for the Viz

library(showtext) # Using Fonts More Easily in R Graphs

library(ggimage) # Using Images in ggplot2

library(fontawesome) # Social Media icons

library(ggtext) # Markdown Text in ggplot2

library(patchwork) # For compiling plots

library(figpatch) # Images in patchwork

library(magick) # Work with Images and Logos

library(ggimage) # Background Image

library(scales) # ggplot2 labels and scaling

#==============================================================================#

# Data Load-in------------------------------------------------------------------

#==============================================================================#

# Option 1: tidytuesdayR package

tuesdata <- tidytuesdayR::tt_load(2024, week = 3)

polling_places <- tuesdata$polling_places

rm(tuesdata)

finaldf <- read_csv(here::here("data", "uspolling.csv"))

plotdf3 <- read_csv(here::here("data", "uspolling1.csv"))

# Data-frame / Tibble for actual visualization

plotdf4 <- plotdf3 |>

filter(

(lat > 0 & lat < 90) &

(long > -180 & long < -67)

) |>

rename(lon = long) |>

usmap::usmap_transform()

df <- polling_places |>

mutate(year = year(election_date)) |>

distinct(year, address)

for (i in seq(2012, 2020, 2)) {

temp <- df |>

filter(year == i) |>

pull(address) |>

unique()

assign_name <- paste0("df_", i)

assign(assign_name, temp)

}

rm(temp)

rm(assign_name)

rm(i)

# A tibble of only polling stations which are new from previous elections

newdf <- df |>

mutate(

new_poll = case_when(

((year == 2014) & !(address %in% df_2012)) ~ TRUE,

((year == 2016) & !(address %in% c(df_2012, df_2014))) ~ TRUE,

((year == 2018) & !(address %in% c(df_2012, df_2014, df_2016))) ~ TRUE,

((year == 2020) & !(address %in% c(df_2012, df_2014, df_2016, df_2018))) ~ TRUE,

.default = FALSE

)

) |>

filter(new_poll) |>

select(-new_poll) |>

count(year)

rm(df_2012)

rm(df_2014)

rm(df_2016)

rm(df_2018)

rm(df_2020)

totaldf <- df |>

count(year)

#==============================================================================#

# Exploratory Data Analysis: Not to Re-Run ------------------------------------

#==============================================================================#

# Aiming to geocode and reverse geocode the entries using

# tidygeocoder R package !

# names of variables in data

names(polling_places)

# Finding number of locations each year

polling_places |>

mutate(year = year(election_date)) |>

distinct(year, address) |>

count(year)

#==============================================================================#

# Data Wrangling : Not to Re-Run -----------------------------------------------

#==============================================================================#

# Find new polling locations for each year

# Restrict analysis to in-person polling stations

polling_places |>

count(location_type)

df <- polling_places |>

mutate(year = year(election_date)) |>

distinct(year, address)

for (i in seq(2012, 2020, 2)) {

temp <- df |>

filter(year == i) |>

pull(address) |>

unique()

assign_name <- paste0("df_", i)

assign(assign_name, temp)

}

rm(temp)

rm(assign_name)

rm(i)

# Saving data on new polling Stations in West Virginia

# Now, using tidygeocoder to get latitude and longitude of the addresses

library(tidygeocoder)

# Using Census Data method for geocoding addresses

plotdf1 <- plotdf |>

geocode(

address = address,

method = "census",

verbose = FALSE)

# Using Open Street Maps to Geocode Addresses

plotdf2 <- plotdf |>

geocode(

address = address,

method = "osm",

verbose = FALSE)

finaldf <- plotdf1 |>

full_join(plotdf2,

by = c("year", "address"),

suffix = c("_df1", "_df2")) |>

mutate(lat = coalesce(lat_df1, lat_df2),

long = coalesce(long_df1, long_df2)) |>

select(-lat_df1, -long_df1, -lat_df2, -long_df2) |>

drop_na()

# To save on computation time, save the final database results in a csv

write_csv(

x = finaldf,

file = here::here("data", "uspolling.csv")

)

# A tibble of only polling stations which are new from previous elections

plotdf <- df |>

mutate(

new_poll = case_when(

((year == 2014) & !(address %in% df_2012)) ~ TRUE,

((year == 2016) & !(address %in% c(df_2012, df_2014))) ~ TRUE,

((year == 2018) & !(address %in% c(df_2012, df_2014, df_2016))) ~ TRUE,

((year == 2020) & !(address %in% c(df_2012, df_2014, df_2016, df_2018))) ~ TRUE,

.default = FALSE

)

) |>

filter(new_poll) |>

select(-new_poll)

# Finding Common Words in Addresses of US Polling Stations

library(tidytext)

common_words <- plotdf |>

unnest_tokens(output = "word",

input = "address") |>

count(word, sort = TRUE) |>

filter(word %in% c("church", "center", "school", "hall"))

finaldf <- plotdf |>

mutate(address = str_to_lower(address)) |>

mutate(

type = case_when(

str_detect(address, "church") ~ "Church",

str_detect(address, "school") ~ "School",

.default = NA

)

) |>

drop_na()

finaldf |>

count(type)

# Now, using tidygeocoder to get latitude and longitude of the addresses

library(tidygeocoder)

# Using Census Data method for geocoding addresses

plotdf1 <- finaldf |>

geocode(

address = address,

method = "census",

verbose = FALSE)

# Using Open Street Maps to Geocode remaining Addresses

plotdf2 <- plotdf1 |>

filter(is.na(lat)) |>

geocode(

address = address,

method = "osm",

verbose = FALSE)

nrow(finaldf)

nrow(plotdf1)

nrow(plotdf2)

plotdf2 <- plotdf2 |>

rename(lat = lat...6,

long = long...7) |>

select(-c(lat...4, long...5))

plotdf1

plotdf2

plotdf3 <- plotdf1 |>

full_join(plotdf2,

by = c("year", "address"),

suffix = c("_df1", "_df2")) |>

mutate(lat = coalesce(lat_df1, lat_df2),

long = coalesce(long_df1, long_df2)) |>

select(-lat_df1, -long_df1, -lat_df2, -long_df2, -type_df2) |>

rename(type = type_df1) |>

drop_na()

plotdf3

# To save on computation time, save the final database results in a csv

write_csv(

x = plotdf3,

file = here::here("data", "uspolling1.csv")

)

plotdf3

#==============================================================================#

# Load curated dataset, cleaned and wrangled -----------------------------------

#==============================================================================#

#==============================================================================#

# Options & Visualization Parameters--------------------------------------------

#==============================================================================#

# Load fonts

font_add_google("Sirin Stencil",

family = "title_font") # Font for titles

font_add_google("Saira Extra Condensed",

family = "caption_font") # Font for the caption

font_add_google("Strait",

family = "body_font") # Font for plot text

showtext_auto()

# Icons to use in graph

# Credits: Used code from

# Creating a Colour Palette for the Visualization

mypal <- paletteer::paletteer_d("nord::victory_bonds")

mypal

# Define colours

school <- "#5773CCFF"

church <- "#FFB900FF"

bg_col <- "#F0F2F8FF" # Background Colour

text_col <- "#010440" # Colour for the text

text_hil <- "#010440" # Colour for highlighted text

# Define Text Size

ts = unit(20, units = "cm") # Text Size

# Caption stuff

sysfonts::font_add(family = "Font Awesome 6 Brands",

regular = here::here("docs", "Font Awesome 6 Brands-Regular-400.otf"))

github <- "&#xf09b"

github_username <- "aditya-dahiya"

xtwitter <- "&#xe61b"

xtwitter_username <- "@adityadahiyaias"

linkedin <- "&#xf08c"

linkedin_username <- "dr-aditya-dahiya-ias"

social_caption <- glue::glue("<span style='font-family:\"Font Awesome 6 Brands\";'>{github};</span> <span style='color: {text_col}'>{github_username} </span> <span style='font-family:\"Font Awesome 6 Brands\";'>{xtwitter};</span> <span style='color: {text_col}'>{xtwitter_username}</span> <span style='font-family:\"Font Awesome 6 Brands\";'>{linkedin};</span> <span style='color: {text_col}'>{linkedin_username}</span>")

# Add text to plot--------------------------------------------------------------

plot_title <- "New Polling Sites in US Elections !"

subtitle_text <- "In the last 5 elections, the United States has introduced over 94,000 new polling sites, most of them in the 2020 elections. Here, we look at these new sites located in Churches or Schools. Most of these new polling sites are concentrated in a few States. Further, more Church-based new polling sites occur in the South-East, while more School-based new polling sites lie in the North-East."

plot_subtitle <- paste(strwrap(subtitle_text, 130), collapse = "\n")

plot_caption <- paste0("**Data & Inspiration:** Center for Public Integrity | ", "**Graphics:** ", social_caption)

#==============================================================================#

# Data Visualization------------------------------------------------------------

#==============================================================================#

g <-

usmap::plot_usmap(

col = "darkgrey",

fill = "white"

) +

geom_point(

data = plotdf4,

mapping = aes(

x = x,

y = y,

col = type,

shape = type),

alpha = 0.8

) +

scale_color_manual(values = c(church, school)) +

labs(

title = plot_title,

subtitle = plot_subtitle,

caption = plot_caption,

color = NULL,

shape = NULL

) +

theme(

legend.position = c(0.7, 0.07),

plot.caption = element_textbox(family = "caption_font",

hjust = 0.5,

colour = text_col,

size = 1.5 * ts),

plot.title = element_text(hjust = 0.5,

size = 6*ts,

family = "title_font",

face = "bold",

colour = text_hil,

margin = margin(4,0,2,0)),

plot.subtitle = element_text(hjust = 0.5,

size = 1.6 * ts,

family = "body_font",

colour = text_col,

margin = margin(5,0,2,0),

lineheight = 0.35),

plot.background = element_rect(fill = bg_col,

color = bg_col,

linewidth = 0),

plot.title.position = "plot",

legend.text = element_text(size = 2 * ts,

family = "body_font",

colour = text_col,

margin = margin(0,0,0,0),

hjust = 0),

legend.key = element_rect(fill = bg_col,

colour = bg_col),

legend.background = element_rect(fill = bg_col)

)

#=============================================================================#

# Image Saving-----------------------------------------------------------------

#=============================================================================#

ggsave(

filename = here::here("docs", "tidy_uspolling.png"),

plot = g,

width = 20,

height = 20,

units = "cm",

bg = bg_col

)