2018 Election Results for Minnesota

ElectionResults2018.R

library("tidyverse")
library("rvest")
library("openxlsx")
library("ggthemes")

# Election Results --------------------------------------------------------

column_names <- c("State",
                  "County_ID",
                  "Precinct_name",
                  "Office_ID",
                  "Office_Name",
                  "District",
                  "Candidate_Order_Code",
                  "Candidate_Name",
                  "Suffix",
                  "Incumbent_Code",
                  "Party_Abbreviation",
                  "Number_of_Precincts_reporting",
                  "Total_number_of_precincts_voting_for_the_office",
                  "Votes_for_Candidate",
                  "Percentage_of_Votes_for_Candidate_out_of_Total_Votes_for_Office",
                  "Total_number_of_votes_for_Office_in_area")

urls <- list()
urls$results <- "https://electionresults.sos.state.mn.us/Results/MediaResult/115?mediafileid=38"
urls$county <- "http://w20.education.state.mn.us/MdeOrgView/reference/county"
urls$wiki <- "https://en.wikipedia.org/wiki/List_of_counties_in_Minnesota"
urls$xlsx <- "https://mn.gov/admin/assets/mn_county_estimates_sdc_2017v2_tcm36-353258.xlsx"

results <- read.csv(urls$results,
                    sep = ";",
                    stringsAsFactors = F,
                    header = F)
names(results) <- column_names

# County Information ------------------------------------------------------

webpage <- read_html(urls$county)
tbls <- html_nodes(webpage, "table")
# Shows all tables
tbls
# Select Table
counties <- html_table(tbls[1],fill = T)[[1]]
names(counties) <- c("County_ID","County_Name")

# County Area Information -------------------------------------------

webpage <- read_html(urls$wiki)
tbls <- html_nodes(webpage, "table")
# Shows all tables
tbls
# Select Table
wiki <- html_table(tbls[1],fill = T)[[1]]
names(wiki) <- gsub("\\[.*$","",names(wiki))
wiki$Population <- wiki$Population %>% 
  gsub("^.*♠","",.) %>%
  gsub("[^0-9\\.]","",.) %>%
  as.numeric
wiki$Area <- wiki$Area %>% 
  gsub("^[0-9]+♠","",.) %>%
  gsub("sq.*$","",.) %>%
  gsub("[^0-9\\.]","",.) %>%
  as.numeric
wiki$Map <- NULL
wiki$County <- gsub(" County","",wiki$County)
names(wiki)[1] <- "County_Name"

# Population Estimates for 2017 -------------------------------------------

estimates <- read.xlsx(urls$xlsx)
# Fix Names (comma + period OR period get replaced with underscore)
names(estimates) <- gsub("(\\,?\\.)","_",names(estimates))
# County IDs are not state codes for the county?
names(estimates)[1] <- "County_ID"
estimates$County_ID <- NULL

# Append County Information to Election Results ---------------------------

counties <- merge(counties, estimates, by = "County_Name", all.x = T) %>%
  merge(wiki, by = "County_Name", all.x = T)
results <- merge(results, counties, by = "County_ID", all.x = T)
results$Density <- results$Total_Population_2017 / results$Area
fix <- which(is.na(results$Density))
results$Density[fix] <- results$Population[fix] / results$Area[fix]
results$Percent_Area <- round((results$Votes_for_Candidate / results$Total_number_of_votes_for_Office_in_area) * 100,digits = 2)

# Senate Results ----------------------------------------------------------

senate <- filter(results, Office_Name == "U.S. Senator" & Party_Abbreviation %in% c("DFL","R"))
senate.hennepin <- filter(senate, County_Name == "Hennepin")
sample <- senate[which(is.na(senate$Density)),]

# Plot Results ------------------------------------------------------------

p1 <- ggplot(senate, aes(x = Density,
                   y = Percent_Area,
                   color = Party_Abbreviation,
                   group = Party_Abbreviation,
                   text = County_Name)) + 
  geom_point() + theme_fivethirtyeight() + 
  scale_y_continuous(limits = c(20,80)) + 
  scale_color_manual(values = c("#2f99ea","#ea2f2f")) +
  scale_x_log10() + 
  geom_density_2d() + facet_wrap(~Party_Abbreviation) + 
  theme(axis.title = element_text(),
        legend.position = "right",
        legend.justification = "top",
        legend.box = "vertical",
        legend.direction = "vertical") +
  labs(x = "\nPeople Per Square Mile (base-10 log scale)",
       caption = "@appupio",
       y = "Percent of Vote\n",
       color = "Party") +
  ggtitle("2018 Election Results for Minnesota")
p1 + geom_smooth()