neodowns Performs computer simulations of candidate competition under different electoral systems (first-past-the-post and ranked-choice voting) based on the spatial model proposed in Atsusaka and Landsman (2024) (working paper).
neodowns can be installed using the following code:
remotes::install_github(
"YukiAtsusaka/neodowns",
dependencies = TRUE
)Our workhorse function sim_data simulates spatial positions of voters and candidates under a set of conditions defined by model parameters. neodowns implements the algorithm based on the neodownsian model of candidate competition.
# Generate simulated data
sim <- sim_data(N_voters = 1000,
N_groups = 3,
dist = "Normal")
library(ggplot2)
voters <- sim$gen_voters
cands <- sim$gen_cands
ggplot() +
geom_point(data = voters,
aes(x, y,
shape = ethnic_group,
color = ethnic_group),
size = 1, alpha = 0.5) +
geom_point(data = cands,
aes(x, y,
shape = ethnic_group,
color = ethnic_group),
size = 4) +
geom_hline(aes(yintercept = 0), color = "black", alpha = 0.5, linetype = "dashed") +
geom_vline(aes(xintercept = 0), color = "black", alpha = 0.5, linetype = "dashed") +
scale_color_manual(
values = c(
alpha("deepskyblue4", 1),
alpha("deeppink4", 1),
alpha("slategray", 1)
),
name = "Group", labels = c("Group A", "Group B", "Group C")
) +
scale_shape_manual(
name = "Group", labels = c("Group A", "Group B", "Group C"),
values = c(16, 17, 15)
) +
theme_minimal() +
theme(legend.position = "none") +
ylab("") +
xlab("")
ggsave("man/fig/example-plot.png", height = 5, width = 5) 
# Implement the algorithm
out <- neodowns(sim, n_iter = 1000)
head(out)
# $voters
# # A tibble: 57,000 × 39
# group ethnic_group x y D1 D2 D3 D4 D5
# <dbl> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1 1 Group 1 0.477 0.426 0.363 1.17 1.36 0.998 1.82
# 2 2 Group 2 -0.369 -0.118 1.14 0.562 0.564 1.91 1.23
# 3 3 Group 3 0.157 -0.547 0.922 1.23 0.486 1.62 1.90
# 4 3 Group 3 -0.213 -0.728 1.31 1.17 0.108 2.02 1.77
# 5 2 Group 2 -1.39 0.953 2.26 0.928 1.96 2.92 0.318
# 6 2 Group 2 -1.39 1.09 2.32 1.03 2.07 2.95 0.452
# 7 1 Group 1 1.34 0.416 0.650 2.03 1.98 0.139 2.68
# 8 3 Group 3 -1.21 -1.80 2.76 2.20 1.43 3.46 2.44
# 9 3 Group 3 -0.797 -1.15 2.02 1.50 0.677 2.74 1.87
# 10 3 Group 3 -0.310 -2.83 3.18 3.20 2.15 3.68 3.62
# $cands
# # A tibble: 360 × 11
# group ethnic_group x y type party dist moderation
# <int> <chr> <dbl> <dbl> <chr> <chr> <dbl> <dbl>
# 1 1 Group 1 0.724 0.194 mod Group_… 0.75 1
# 2 2 Group 2 -0.678 0.321 mod Group_… 0.75 1
# 3 3 Group 3 -0.288 -0.692 mod Group_… 0.75 1
# 4 1 Group 1 1.45 0.388 ext Group_… 1.5 1
# 5 2 Group 2 -1.36 0.642 ext Group_… 1.5 1
# 6 3 Group 3 -0.577 -1.38 ext Group_… 1.5 1
# 7 1 Group 1 0.766 0.212 mod Group_… 0.755 1.05
# 8 2 Group 2 -0.652 0.311 mod Group_… 0.772 0.936
# 9 3 Group 3 -0.290 -0.725 mod Group_… 0.747 1.05
# 10 1 Group 1 1.52 0.391 ext Group_… 1.52 1.03 # Extract quantities of interest
extract(out, quantity = "ideology")
# # A tibble: 60 × 3
# # Groups: system [3]
# system iter `mean(moderation)`
# <chr> <dbl> <dbl>
# 1 max1 1 1
# 2 max1 2 1.00
# 3 max1 3 1.01
# 4 max1 4 0.959
# 5 max1 5 0.974
# 6 max1 6 0.931
# 7 max1 7 0.932
# 8 max1 8 0.907
# 9 max1 9 0.900
# 10 max1 10 0.882
extract(out, quantity = "ethnic")
# # A tibble: 57,000 × 4
# iter ethnic_group ethnic_voting_first ethnic_voting_second
# <int> <chr> <dbl> <dbl>
# 1 2 Group 1 0.949 NA
# 2 2 Group 2 0.707 NA
# 3 2 Group 3 0.831 NA
# 4 2 Group 3 0.749 NA
# 5 2 Group 2 0.991 NA
# 6 2 Group 2 0.894 NA
# 7 2 Group 1 0.908 NA
# 8 2 Group 3 0.966 NA
# 9 2 Group 3 0.818 NA
# 10 2 Group 3 0.969 NA