-
Notifications
You must be signed in to change notification settings - Fork 0
/
visualizations.R
192 lines (169 loc) · 9.05 KB
/
visualizations.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
#laina emmons 4/10/2021
#regression analysis on States & Counties dataset
#dot-and-whisker visualization of regression tables
setwd("/Users/lainaemmons/Documents/GitHub/senior-thesis")
getwd()
library(dplyr)
library(tidytext)
library(ggplot2)
library(tidyr)
library(stringr)
library(broom)
data <- read.csv("states_and_counties.csv")
#colnames(data)
data = subset(data, select = -c(inf_mar, inf_apr, inf_may, inf_jun, inf_jul, inf_aug, Demographics,
inf_sep_13, death_mar, death_apr, death_may, death_jun, death_jul,
death_aug, death_sep, med_income, num_polls, Voting, Election,
difference_turnout_2020_minus_2016, difference_trump_perc_2020_minus_2016,
difference_gop_gov_perc_2020_minus_2016, X2020_registered, X2016_registered))
data$inf_by_pop <- as.numeric(as.character(data$inf_by_pop))
data$death_rate_by_pop <- as.numeric(as.character(data$death_rate_by_pop))
data$perc_non_white <-as.numeric(as.character(data$perc_non_white))
data$perc_below_poverty_level <- as.numeric(as.character(data$perc_below_poverty_level))
data$unemployment <- as.numeric(as.character(data$unemployment))
data$polls_per_person <- as.numeric(as.character(data$polls_per_person))
data$early_vote_duration <- as.numeric(as.character(data$early_vote_duration))
data$perc_trump_2016 <- as.numeric(as.character(data$perc_trump_2016))
data$perc_trump_2020 <- as.numeric(as.character(data$perc_trump_2020))
data$perc_rep_gov_2016 <- as.numeric(as.character(data$perc_rep_gov_2016))
data$perc_rep_gov_2020 <- as.numeric(as.character(data$perc_trump_2020))
data$X2016_turnout <- as.numeric(as.character(data$X2016_turnout))
data$X2020_turnout <- as.numeric(as.character(data$X2020_turnout))
colnames(data)
data <- data %>%
filter(!death_rate_by_pop == "NA")
data <- data %>%
filter(!inf_by_pop == "NA")
data <- data %>%
filter(!unemployment == "NA")
#data <- data %>%
# filter(!gov_rep_incumbent_2020 == 0)
summary(data)
###### Turnout?
turnout <- data %>%
filter(gov_rep_incumbent_2020 == 1)
turnout <- lm(X2020_turnout ~ death_rate_by_pop + inf_by_pop + perc_non_white +
unemployment + polls_per_person + early_vote_duration + same_day_reg +
vra_provision + X2016_turnout, data = turnout)
summary(turnout)
turnout
dwplot(turnout, show_intercept = FALSE, ci = .95) %>%
relabel_predictors(c(inf_by_pop = "Infection Rate",
perc_non_white = "% Population Non-White",
unemployment = "Unemployment Rate",
polls_per_person = "Polls per Capita",
early_vote_duration = "Duration of Early Vote Period",
same_day_reg = "Offers Same-Day Registration",
vra_provision = "Subject to VRA Provision",
perc_trump_2016 = "Trump 2016 %",
perc_rep_gov_2016 = "GOP Govornor 2016 %")) +
ylab("Predictors") + xlab("Coefficient Estimate") +
ggtitle("Regression Analysis on 2020 Election Models") +
theme_bw() +
theme(axis.text = element_text(color = "black", face = "bold")) +
labs(subtitle = "by Laina Emmons") +
scale_color_brewer(palette = "Set1",
breaks = c("Trump Share of Vote", "GOP Gov Share of Vote")) +
theme(plot.title = element_text(face="bold"),
plot.subtitle = element_text(face = "bold"),
legend.position = c(0.007, 0.01),
legend.justification = c(0, 0),
legend.background = element_rect(colour="grey80"),
legend.title = element_blank())
###### deaths only dotwhisker regression plots
library(dotwhisker)
gub_inc <- data %>%
filter(gov_rep_incumbent_2020 == 1)
gub_inc <- lm(perc_rep_gov_2020 ~ death_rate_by_pop + perc_non_white +
unemployment + polls_per_person + early_vote_duration + same_day_reg +
vra_provision + perc_rep_gov_2016, data = gub_inc)
summary(gub_inc)
trump_inc <- data %>%
filter(gov_rep_incumbent_2020 == 1)
trump_inc <- lm(perc_trump_2020 ~ death_rate_by_pop + perc_non_white +
unemployment + polls_per_person + early_vote_duration + same_day_reg +
vra_provision + perc_trump_2016, data = trump_inc)
summary(trump_inc)
death_model <- bind_rows(
tidy(gub_inc) %>% mutate(model = "GOP Gov Share of Vote"),
tidy(trump_inc) %>% mutate(model = "Trump Share of Vote"))
death_model$model <- factor(death_model$model,
levels = c("Trump Share of Vote", "GOP Gov Share of Vote"),
labels = c("Trump Share of Vote", "GOP Gov Share of Vote")
)
dwplot(death_model, show_intercept = FALSE, ci = .95) %>%
relabel_predictors(c(death_rate_by_pop = "Death Rate",
perc_non_white = "% Population Non-White",
unemployment = "Unemployment Rate",
polls_per_person = "Polls per Capita",
early_vote_duration = "Duration of Early Vote Period",
same_day_reg = "Offers Same-Day Registration",
vra_provision = "Subject to VRA Provision",
perc_trump_2016 = "Trump 2016 %",
perc_rep_gov_2016 = "GOP Govornor 2016 %")) +
ylab("Predictors") + xlab("Coefficient Estimate") +
ggtitle("Regression Analysis on 2020 Election Models") +
theme_bw() +
labs(subtitle = " This is a dot-and-whisker plot, the dots represent the regression coefficient estimate and the whiskers represent confidence interval.
This plot shows that the variable is statistically relevant if the whisker does not cross 0.
If the coefficient estimate > 0, the variable positively impacted the share of votes, and vice versa.
by Laina Emmons") +
scale_color_brewer(palette = "Set1",
breaks = c("Trump Share of Vote", "GOP Gov Share of Vote")) +
theme(plot.title = element_text(face="bold"),
plot.subtitle = element_text(face = "bold"),
legend.position = c(0.007, 0.01),
legend.justification = c(0, 0),
legend.background = element_rect(colour="grey80"),
legend.title = element_blank())
##### infections only dotwhisker
gub_inc_gov <- data %>%
filter(gov_rep_incumbent_2020 == 1)
gub_inc_gov <- lm(perc_rep_gov_2020 ~ inf_by_pop + perc_non_white + unemployment +
polls_per_person + early_vote_duration + same_day_reg +
vra_provision + perc_rep_gov_2016, data = gub_inc_gov)
summary(gub_inc_gov)
trump_inc_gov <- data %>%
filter(gov_rep_incumbent_2020 == 1)
trump_inc_gov <- lm(perc_trump_2020 ~ inf_by_pop + perc_non_white +
unemployment + polls_per_person + early_vote_duration + same_day_reg +
vra_provision + perc_trump_2016, data = trump_inc_gov)
summary(trump_inc_gov)
inf_model <- bind_rows(
tidy(gub_inc_gov) %>% mutate(model = "GOP Gov Share of Vote"),
tidy(trump_inc_gov) %>% mutate(model = "Trump Share of Vote"))
inf_model$model <- factor(inf_model$model,
levels = c("Trump Share of Vote", "GOP Gov Share of Vote"),
labels = c("Trump Share of Vote", "GOP Gov Share of Vote"))
dwplot(inf_model, show_intercept = FALSE, ci = .95) %>%
relabel_predictors(c(inf_by_pop = "Infection Rate",
perc_non_white = "% Population Non-White",
unemployment = "Unemployment Rate",
polls_per_person = "Polls per Capita",
early_vote_duration = "Duration of Early Vote Period",
same_day_reg = "Offers Same-Day Registration",
vra_provision = "Subject to VRA Provision",
perc_trump_2016 = "Trump 2016 %",
perc_rep_gov_2016 = "GOP Govornor 2016 %")) +
ylab("Predictors") + xlab("Coefficient Estimate") +
ggtitle("Regression Analysis on 2020 Election Models") +
theme_bw() +
labs(subtitle = " This is a dot-and-whisker plot, the dots represent the regression coefficient estimate and the whiskers represent confidence interval.
This plot shows that the variable is statistically relevant if the whisker does not cross 0.
If the coefficient estimate > 0, the variable positively impacted the share of votes, and vice versa.
by Laina Emmons") +
scale_color_brewer(palette = "Set1",
breaks = c("Trump Share of Vote", "GOP Gov Share of Vote")) +
theme(plot.title = element_text(face="bold"),
plot.subtitle = element_text(face = "bold"),
legend.position = c(0.007, 0.01),
legend.justification = c(0, 0),
legend.background = element_rect(colour="grey80"),
legend.title = element_blank())
#### compared to democrat incumbents
dems <- data %>%
filter(State == c("Delaware", "North Carolina", "Washington"))
dem_reg <- lm(perc_trump_2020 ~ inf_by_pop + death_rate_by_pop + perc_non_white +
unemployment + polls_per_person + early_vote_duration + same_day_reg +
vra_provision + perc_trump_2016, data = dems)
summary(dem_reg)