This document presents the analysis and data processing steps for the Election Campaign Simulation. The focus is on precinct-level election data and voter files to identify GOTV (Get Out The Vote) and persuasion targets.
library(tidyverse)
library(knitr)data_2016 <- read_delim("C:/Users/alesa/OneDrive/Desktop/grad school/POS 6933/Election Simulation/Data/Precinct Results/DAD_PctResults20161108.txt", col_names = FALSE, col_select = c(6, 12, 16, 19))
colnames(data_2016) <- c("precinct", "contest_name", "candidate_party", "vote_total")
data_2018 <- read_delim("C:/Users/alesa/OneDrive/Desktop/grad school/POS 6933/Election Simulation/Data/Precinct Results/DAD_PctResults20181106.txt", col_names = FALSE, col_select = c(6, 12, 16, 19))
colnames(data_2018) <- c("precinct", "contest_name", "candidate_party", "vote_total")
data_2020 <- read_delim("C:/Users/alesa/OneDrive/Desktop/grad school/POS 6933/Election Simulation/Data/Precinct Results/DAD_PctResults20201103.txt", col_names = FALSE, col_select = c(6, 12, 16, 19))
colnames(data_2020) <- c("precinct", "contest_name", "candidate_party", "vote_total")
data_2022 <- read_delim("C:/Users/alesa/OneDrive/Desktop/grad school/POS 6933/Election Simulation/Data/Precinct Results/DAD_PctResults20221108.txt", col_names = FALSE, col_select = c(6, 12, 16, 19))
colnames(data_2022) <- c("precinct", "contest_name", "candidate_party", "vote_total")data_2018 <- data_2018 %>% mutate(precinct = str_pad(precinct, 3, pad = "0"))
data_2020 <- data_2020 %>% mutate(precinct = str_sub(precinct, 1, 3))
data_2022 <- data_2022 %>% mutate(precinct = str_sub(precinct, 1, 3))data_2022 <- data_2022 %>% mutate(contest_name = ifelse(contest_name == "Governor and Lieutenant Governor", "Governor", contest_name))data_2016_processed <- data_2016 %>%
filter(contest_name %in% c("President of the United States", "Representative in Congress", "United States Senator", "Governor", "State Representative", "State Senator")) %>%
group_by(precinct, contest_name) %>%
summarize(
dem_votes = sum(vote_total[candidate_party == "DEM"]),
rep_votes = sum(vote_total[candidate_party == "REP"]),
total_votes = dem_votes + rep_votes,
vote_share_dem = dem_votes / total_votes * 100,
vote_share_rep = rep_votes / total_votes * 100
) %>%
ungroup() %>%
mutate(year = 2016) %>%
select(precinct, contest_name, year, vote_share_dem, vote_share_rep, total_votes) %>%
filter(total_votes >= 100)
data_2018_processed <- data_2018 %>%
filter(contest_name %in% c("President of the United States", "Representative in Congress", "United States Senator", "Governor", "State Representative", "State Senator")) %>%
group_by(precinct, contest_name) %>%
summarize(
dem_votes = sum(vote_total[candidate_party == "DEM"]),
rep_votes = sum(vote_total[candidate_party == "REP"]),
total_votes = dem_votes + rep_votes,
vote_share_dem = dem_votes / total_votes * 100,
vote_share_rep = rep_votes / total_votes * 100
) %>%
ungroup() %>%
mutate(year = 2018) %>%
select(precinct, contest_name, year, vote_share_dem, vote_share_rep, total_votes) %>%
filter(total_votes >= 100)
data_2020_processed <- data_2020 %>%
filter(contest_name %in% c("President of the United States", "Representative in Congress", "United States Senator", "Governor", "State Representative", "State Senator")) %>%
group_by(precinct, contest_name) %>%
summarize(
dem_votes = sum(vote_total[candidate_party == "DEM"]),
rep_votes = sum(vote_total[candidate_party == "REP"]),
total_votes = dem_votes + rep_votes,
vote_share_dem = dem_votes / total_votes * 100,
vote_share_rep = rep_votes / total_votes * 100
) %>%
ungroup() %>%
mutate(year = 2020) %>%
select(precinct, contest_name, year, vote_share_dem, vote_share_rep, total_votes) %>%
filter(total_votes >= 100)
data_2022_processed <- data_2022 %>%
filter(contest_name %in% c("President of the United States", "Representative in Congress", "United States Senator", "Governor", "State Representative", "State Senator")) %>%
group_by(precinct, contest_name) %>%
summarize(
dem_votes = sum(vote_total[candidate_party == "DEM"]),
rep_votes = sum(vote_total[candidate_party == "REP"]),
total_votes = dem_votes + rep_votes,
vote_share_dem = dem_votes / total_votes * 100,
vote_share_rep = rep_votes / total_votes * 100
) %>%
ungroup() %>%
mutate(year = 2022) %>%
select(precinct, contest_name, year, vote_share_dem, vote_share_rep, total_votes) %>%
filter(total_votes >= 100)combined_data <- bind_rows(data_2016_processed, data_2018_processed, data_2020_processed, data_2022_processed)average_and_sd <- combined_data %>%
group_by(precinct) %>%
summarize(
avg_vote_share_dem = mean(vote_share_dem, na.rm = TRUE),
sd_vote_share_dem = sd(vote_share_dem, na.rm = TRUE)
)gotv_target_precincts <- average_and_sd %>%
arrange(desc(avg_vote_share_dem)) %>%
slice_head(n = 25)| Precinct- | Dem Vote Share- | Vote Variation |
|---|---|---|
| 222 | 95.57956 | 1.350201 |
| 172 | 95.37798 | 2.949766 |
| 508 | 95.16253 | 3.238699 |
| 216 | 93.99078 | 2.530406 |
| 251 | 93.75685 | 3.565219 |
| 221 | 93.72252 | 2.520116 |
| 173 | 93.63684 | 4.108892 |
| 257 | 93.54195 | 3.282334 |
| 218 | 93.53312 | 3.430555 |
| 520 | 93.51469 | 3.085034 |
| 214 | 93.40552 | 2.250026 |
| 511 | 93.36604 | 3.622358 |
| 205 | 93.34540 | 3.238355 |
| 505 | 93.28404 | 4.008555 |
| 145 | 93.26859 | 3.324025 |
| 269 | 93.13085 | 2.718910 |
| 506 | 93.10736 | 4.299125 |
| 217 | 92.98806 | 2.657165 |
| 254 | 92.93799 | 4.164430 |
| 521 | 92.88133 | 4.514028 |
| 253 | 92.87898 | 3.727270 |
| 507 | 92.87455 | 4.384368 |
| 519 | 92.65620 | 4.167402 |
| 213 | 92.59933 | 2.512293 |
| 206 | 92.44235 | 2.838327 |
persuasion_target_precincts <- average_and_sd %>%
arrange(desc(sd_vote_share_dem)) %>%
slice_head(n = 25)| Precinct- | Dem Vote Share- | Vote Variation |
|---|---|---|
| 116 | 44.14024 | 18.68507 |
| 003 | 39.55866 | 18.67399 |
| 067 | 41.69257 | 18.40957 |
| 001 | 37.91634 | 18.21702 |
| 002 | 45.30230 | 17.70280 |
| 004 | 51.89023 | 17.57602 |
| 006 | 41.87254 | 17.17492 |
| 120 | 50.99085 | 16.48928 |
| 181 | 63.43092 | 16.47919 |
| 005 | 47.95894 | 16.34286 |
| 892 | 46.95068 | 16.16370 |
| 119 | 54.70992 | 15.89984 |
| 981 | 60.69190 | 15.79222 |
| 124 | 54.34991 | 15.62646 |
| 833 | 49.69262 | 15.43914 |
| 980 | 63.25100 | 15.01594 |
| 527 | 60.21388 | 14.90112 |
| 025 | 46.95226 | 14.70331 |
| 114 | 51.86225 | 14.47704 |
| 099 | 50.73125 | 14.38135 |
| 157 | 68.08453 | 14.34629 |
| 109 | 58.32753 | 14.25152 |
| 009 | 53.88160 | 14.24967 |
| 028 | 48.78921 | 14.21243 |
| 201 | 52.97533 | 14.06964 |
voter_file <- read_delim("C:/Users/alesa/OneDrive/Desktop/grad school/POS 6933/Election Simulation/Data/VF/DAD_20240924.txt", col_names = FALSE, col_select = c(2, 3, 4, 5, 6, 8, 9, 10, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 29, 35, 36, 37, 38))
colnames(voter_file) <- c("voter_id", "name_last", "name_suffix", "name_first", "name_middle", "res_address_1", "res_address_2", "city", "zipcode", "mail_address_1", "mail_address_2", "mail_address_3", "mail_city", "mail_state", "mail_zip", "mail_country", "gender", "race", "birth_date", "reg_date", "party", "precinct", "status", "area_code", "phone_number", "phone_ext", "email")gotv_voters <- voter_file %>%
filter(precinct %in% gotv_target_precincts$precinct, party == "DEM")
persuasion_voters <- voter_file %>%
filter(precinct %in% persuasion_target_precincts$precinct, party == "NPA")voter_history <- read_delim("C:/Users/alesa/OneDrive/Desktop/grad school/POS 6933/Election Simulation/Data/VF/DAD_H_20210112.txt", col_names = FALSE, col_select = c(2, 4, 5))
colnames(voter_history) <- c("voter_id", "election_type", "history_code")
ballots_not_accepted <- voter_history %>%
filter(history_code %in% c("B", "L")) %>%
filter(election_type %in% c("GEN"))
gotv_targets_not_accepted <- gotv_voters %>%
inner_join(ballots_not_accepted, by = "voter_id")write_csv(gotv_voters, "gotv_voter_contacts.csv")
write_csv(persuasion_voters, "persuasion_voter_contacts.csv")
write_csv(gotv_targets_not_accepted, "vote_by_mail_reminders.csv")This analysis successfully identified target precincts and voter groups for GOTV and persuasion efforts. Additionally, specialized lists were created for micro-targeting efforts, such as reminding voters whose ballots were not accepted to participate in upcoming elections.