세종대왕의 세법개혁과 국민투표 (ggplot)
Data
원자료는 세종실록, 요약표는 오기수 교수의 논문에서.
| 지역 | 계급 | 신분 | 찬반 | 집계 |
|---|---|---|---|---|
| 서울 | 대신 등 | 관료 | 찬성 | 21 |
| 서울 | 대신 등 | 관료 | 반대 | 194 |
| 서울 | 3품이하현직 | 관료 | 찬성 | 259 |
| 서울 | 3품이하현직 | 관료 | 반대 | 393 |
| 서울 | 3품이하전직 | 관료 | 찬성 | 443 |
| 서울 | 3품이하전직 | 관료 | 반대 | 117 |
| 유후사 | 품관촌민 | 품관촌민 | 찬성 | 1123 |
| 유후사 | 품관촌민 | 품관촌민 | 반대 | 71 |
| 경기 | 수령 | 관료 | 찬성 | 29 |
| 경기 | 수령 | 관료 | 반대 | 5 |
| 경기 | 품관촌민 | 품관촌민 | 찬성 | 17076 |
| 경기 | 품관촌민 | 품관촌민 | 반대 | 236 |
| 평안 | 대신 등 | 관료 | 반대 | 1 |
| 평안 | 수령 | 관료 | 찬성 | 6 |
| 평안 | 수령 | 관료 | 반대 | 35 |
| 평안 | 품관촌민 | 품관촌민 | 찬성 | 1326 |
| 평안 | 품관촌민 | 품관촌민 | 반대 | 28474 |
| 황해 | 수령 | 관료 | 찬성 | 17 |
| 황해 | 수령 | 관료 | 반대 | 17 |
| 황해 | 품관촌민 | 품관촌민 | 찬성 | 4454 |
| 황해 | 품관촌민 | 품관촌민 | 반대 | 15601 |
| 충청 | 대신 등 | 관료 | 반대 | 2 |
| 충청 | 수령 | 관료 | 찬성 | 35 |
| 충청 | 수령 | 관료 | 반대 | 26 |
| 충청 | 품관촌민 | 품관촌민 | 찬성 | 6982 |
| 충청 | 품관촌민 | 품관촌민 | 반대 | 14013 |
| 강원 | 수령 | 관료 | 찬성 | 5 |
| 강원 | 수령 | 관료 | 반대 | 10 |
| 강원 | 품관촌민 | 품관촌민 | 찬성 | 939 |
| 강원 | 품관촌민 | 품관촌민 | 반대 | 6888 |
| 함길 | 대신 등 | 관료 | 반대 | 1 |
| 함길 | 수령 | 관료 | 찬성 | 3 |
| 함길 | 수령 | 관료 | 반대 | 14 |
| 함길 | 품관촌민 | 품관촌민 | 찬성 | 75 |
| 함길 | 품관촌민 | 품관촌민 | 반대 | 7387 |
| 경상 | 수령 | 관료 | 찬성 | 55 |
| 경상 | 수령 | 관료 | 반대 | 16 |
| 경상 | 품관촌민 | 품관촌민 | 찬성 | 36262 |
| 경상 | 품관촌민 | 품관촌민 | 반대 | 377 |
| 전라 | 대신 등 | 관료 | 반대 | 2 |
| 전라 | 수령 | 관료 | 찬성 | 42 |
| 전라 | 수령 | 관료 | 반대 | 12 |
| 전라 | 품관촌민 | 품관촌민 | 찬성 | 29505 |
| 전라 | 품관촌민 | 품관촌민 | 반대 | 257 |
전체 찬반
vote_total %>%
as.matrix %>%
t %>%
format(big.mark = ",") %>%
kable(caption = "전체 찬반", align = "r")| 찬성 | 반대 |
|---|---|
| 98,657 | 74,149 |
#> `as.matrix`를 `as.table`로 바꿔서 수행하여도 됨.
vote_total %>%
as.matrix %>%
t %>%
prop.table %>%
`*`(100) %>%
format(digits = 3, nsmall = 1) %>%
kable(caption = "전체 찬반(%)", align = "r")| 찬성 | 반대 |
|---|---|
| 57.1 | 42.9 |
원형그래프
source("theme_kr.R")
# str(theme_kr)
# str(vote_total)
vote_total %>% str## 'xtabs' int [1:2(1d)] 98657 74149
## - attr(*, "dimnames")=List of 1
## ..$ vote: chr [1:2] "찬성" "반대"
## - attr(*, "call")= language xtabs(formula = counts ~ vote, data = .)# vote_df <- as.data.frame(vote_total)
vote_df <- vote_total %>% as.data.frame
# n_vote <- length(levels(vote_df[, 1]))
n_vote <- vote_df %>% `[`(, 1) %>% levels %>% length
# y_coord <- vote_df$Freq/2 + c(0, cumsum(head(vote_df$Freq, -1)))
y_coord <- vote_df %>% `[`(, "Freq") %>% `/`(2) %>% `+`(., vote_df %>% `[`(, "Freq") %>% head(n = -1) %>% c(0, .))
# pie_label <- paste(levels(vote_df$vote),
# format(vote_df$Freq, big.mark = ","), sep = "\n")
pie_label <- vote_df %>% .$Freq %>% levels %>%
paste(., vote_df %>% .$Freq %>% format(big.mark = ","), sep = "\n")
pie_label## [1] "\n98,657" "\n74,149"# kable(vote_df, align = "r")
vote_df %>% kable(align = "r")| vote | Freq |
|---|---|
| 찬성 | 98657 |
| 반대 | 74149 |
# str(vote_df)
vote_df %>% str## 'data.frame': 2 obs. of 2 variables:
## $ vote: Factor w/ 2 levels "찬성","반대": 1 2
## $ Freq: int 98657 74149
# p1 <- ggplot(data = vote_df, mapping = aes(x = "", y = Freq, fill = vote))
p1 <- ggplot(data = vote_df, mapping = aes(x = "", y = vote_df[, 2], fill = vote_df[, 1]))
p2 <- p1 +
geom_bar(width = 1, stat = "identity",
position = position_stack(reverse = TRUE))
p2
p2_2 <- p1 +
geom_bar(width = 1, stat = "identity", position = "dodge")
p2_2
p3_2 <- p2_2 +
theme_bw() +
theme_kr
p3_2
p4_2 <- p3_2 +
scale_x_discrete(name = "찬반") +
scale_y_continuous(name = "집계", breaks = vote_df[, 2],
labels = format(vote_df[, 2], big.mark = ","))
p4_2
p5_2 <- p4_2 +
scale_fill_manual(name = "찬반",
values = rainbow(n_vote)[n_vote:1],
labels = vote_df$vote)
p5_2
p6_2 <- p5_2 +
geom_text(aes(y = vote_df[, 2] / 2),
label = format(rev(vote_df[, 2]), big.mark = ","),
# family = "KoPubWorldDotum Medium",
position = position_dodge(width = 1))
p6_2
p7_2 <- p6_2 +
theme(legend.position = c(0.25, 0.75))
p7_2
p8_2 <- p5_2 +
guides(fill = "none") +
geom_text(aes(y = vote_df[, 2] / 2),
family = "KoPubWorldDotum Medium",
label = rev(pie_label),
position = position_dodge(width = 1))
p8_2
# ggsave("../pics/sejong_geom_bar_total_ggplot.png", p7_2, dpi = 72)
p3 <- p2 + theme_bw()
# + theme_kr
p3 
p4 <- p3 +
scale_x_discrete(name = "찬반") +
scale_y_continuous(name = "집계",
breaks = cumsum(vote_df[, 2]),
labels = format(cumsum(vote_df[, 2]), big.mark = ","))
p4
p5 <- p4 +
scale_fill_manual(name = "찬반",
values = rainbow(2)[2:1], labels = vote_df[, 1])
p5
p6 <- p5 +
geom_text(aes(y = y_coord),
label = format(vote_df[, 2], big.mark = ","),
# family = "KoPubWorldDotum Medium",
position = position_identity())
p6
# ggsave("../pics/sejong_geom_bar_total_ggplot_stack.png", p6, dpi = 72)
p7 <- p6 +
theme(legend.position = c(0.2, 0.75))
p7
p8 <- p5 +
guides(fill = "none") +
geom_text(aes(y = y_coord),
label = pie_label,
family = "KoPubWorldDotum Medium",
position = position_identity())
p8
pie_total_1 <- p2 +
coord_polar(theta = "y",
start = 3 * pi /2 ,
direction = -1)
pie_total_1
pie_total_2 <- pie_total_1 +
scale_y_continuous(name = "",
breaks = NULL) +
scale_x_discrete(name = "")
pie_total_2
pie_total_3 <- pie_total_2 +
scale_fill_manual(name = "",
values = rainbow(n_vote)[n_vote:1])
pie_total_3
pie_total_4 <- pie_total_3 +
theme_void(base_family = "KoPubWorldDotum Medium")
pie_total_4
pie_total_5 <- pie_total_4 +
guides(fill = "none")
pie_total_5
pie_total_6 <- pie_total_5 +
geom_text(aes(y = y_coord),
label = pie_label,
family = "KoPubWorldDotum Medium",
position = position_identity())
pie_total_6
pie_total_7 <- pie_total_6 +
ggtitle("전체 찬반") +
theme(plot.margin = unit(c(1, 1, 1.5, 1), "lines"),
plot.title = element_text(hjust = 0.5, size = 20))
pie_total_7
ggsave("../pics/sejong_total_pie_ggplot.png", pie_total_7, dpi = 72)
pie_gg()
pie_gg <- function(df, ggtitle = "", font_family = "KoPubWorldDotum Medium"){
n <- length(names(df))
y_coord <- df[, 2] / 2 + c(0, cumsum(head(df[, 2], -1)))
pie_label <- paste(levels(df[, 1]), format(df[, 2], big.mark = ","),
sep = "\n")
p1 <- ggplot(df, aes(x = "",
y = df[, 2],
fill = df[, 1]))
p2 <- p1 +
geom_bar(width = 1,
stat = "identity",
position = position_stack(reverse = TRUE))
pie_1 <- p2 +
coord_polar(theta = "y",
start = 3 * pi / 2,
direction = -1)
pie_2 <- pie_1 +
scale_y_continuous(name = "",
breaks = NULL) +
scale_x_discrete(name = "")
pie_3 <- pie_2 +
scale_fill_manual(name = "",
values = rainbow(n)[n:1])
pie_4 <- pie_3 +
theme_void(base_family = font_family)
pie_5 <- pie_4 +
guides(fill = "none")
pie_6 <- pie_5 +
geom_text(aes(y = y_coord),
label = pie_label,
family = font_family,
position = position_identity())
pie_7 <- pie_6 +
ggtitle(ggtitle) +
theme(plot.margin = unit(c(1, 1, 1.5, 1), "lines"),
plot.title = element_text(hjust = 0.5))
return(pie_7)
}
dump("pie_gg",
file = "./pie_gg.R")
pie_gg(vote_df,
ggtitle = "전체 찬반",
font_family = "KoPubWorldDotum Medium")
계급 및 지역별 찬반
계급별 찬반
vote_class %>%
format(big.mark = ",") %>%
kable(align = "r", caption = "계급별 찬반")| 대신 등 | 3품이하현직 | 3품이하전직 | 수령 | 품관촌민 | |
|---|---|---|---|---|---|
| 찬성 | 21 | 259 | 443 | 192 | 97,742 |
| 반대 | 200 | 393 | 117 | 135 | 73,304 |
# kable(format(vote_class,
# big.mark = ","),
# align = "r",
# caption = "계급별 찬반")
vote_class %>%
prop.table(margin = 2) %>%
`*`(100) %>%
format(digits = 3, nsmall = 1) %>%
kable(align = "r", caption = "계급별 찬반(%)")| 대신 등 | 3품이하현직 | 3품이하전직 | 수령 | 품관촌민 | |
|---|---|---|---|---|---|
| 찬성 | 9.5 | 39.7 | 79.1 | 58.7 | 57.1 |
| 반대 | 90.5 | 60.3 | 20.9 | 41.3 | 42.9 |
# kable(format(prop.table(vote_class,
# margin = 2) * 100,
# digits = 3,
# nsmall = 1),
# align = "r",
# caption = "계급별 찬반(%)")품관촌민 별도 분석
품관촌민의 수효가 상대적으로 많아서
vote_class_2 %>%
format(big.mark = ",") %>%
kable(align = rep("r", 2), caption = "품관촌민")| 관료 | 품관촌민 | |
|---|---|---|
| 찬성 | 915 | 97,742 |
| 반대 | 845 | 73,304 |
# kable(format(vote_class_2,
# big.mark = ","),
# align = rep("r", 2),
# caption = "품관촌민")소계를 교차표의 주변에 계산
vote_class_2 %>%
cbind("계" = rowSums(.)) %>%
rbind("계" = colSums(.)) %>%
format(big.mark = ",") %>%
kable(caption = "마진 합 추가", align = "r") | 관료 | 품관촌민 | 계 | |
|---|---|---|---|
| 찬성 | 915 | 97,742 | 98,657 |
| 반대 | 845 | 73,304 | 74,149 |
| 계 | 1,760 | 171,046 | 172,806 |
백분율을 계산하여 주변에 집계.
vote_class_2 %>%
prop.table(margin = 2) %>%
`*`(100) %>%
format(digits = 3, nsmall = 1) %>%
kable(align = "r", caption = "관료와 품관촌민(%)")| 관료 | 품관촌민 | |
|---|---|---|
| 찬성 | 52.0 | 57.1 |
| 반대 | 48.0 | 42.9 |
# kable(format(prop.table(vote_class_2,
# margin = 2) * 100,
# digits = 3, nsmall = 1),
# caption = "관료와 품관촌민(%)",
# align = "r")data frame 변환
vote_class_2_df <- as.data.frame(vote_class_2)
vnames_class_kr <- c("찬반", "계급", "집계")
vote_class_2_df %>%
format(big.mark = ",") %>%
kable(align = c('c', 'c', 'r'),
col.names = vnames_class_kr,
caption = "관료와 품관촌민")| 찬반 | 계급 | 집계 |
|---|---|---|
| 찬성 | 관료 | 915 |
| 반대 | 관료 | 845 |
| 찬성 | 품관촌민 | 97,742 |
| 반대 | 품관촌민 | 73,304 |
# kable(format(vote_class_2_df,
# big.mark = ","),
# align = c('c', 'c', 'r'),
# col.names = vnames_class_kr,
# caption = "관료와 품관촌민")
vote_bureaus_df <- subset(vote_class_2_df,
vote_class_2_df$class_2 == "관료",
select = c("vote", "Freq"))
vote_bureaus_df %>%
format(big.mark = ",") %>%
kable(align = 'r',
col.names = vnames_class_kr[-2],
caption = "관료의 찬반")| 찬반 | 집계 |
|---|---|
| 찬성 | 915 |
| 반대 | 845 |
# kable(format(vote_bureaus_df,
# big.mark = ","),
# align = 'r',
# col.names = vnames_class_kr[-2],
# caption = "관료의 찬반")
vote_commons_df <- subset(vote_class_2_df,
vote_class_2_df$class_2 == "품관촌민",
select = c("vote", "Freq"))
vote_commons_df %>%
format(big.mark = ",") %>%
kable(align = 'r',
row.names = FALSE,
col.names = vnames_class_kr[-2],
caption = "품관촌민의 찬반")| 찬반 | 집계 |
|---|---|
| 찬성 | 97,742 |
| 반대 | 73,304 |
# kable(format(vote_commons_df,
# big.mark = ","),
# align = 'r',
# row.names = FALSE,
# col.names = vnames_class_kr[-2],
# caption = "품관촌민의 찬반")
원형그래프
pie_bureaus <- pie_gg(vote_bureaus_df,
ggtitle = "관료의 찬반",
font_family = "KoPubWorldDotum Medium")
pie_commons <- pie_gg(vote_commons_df,
ggtitle = "품관촌민의 찬반",
font_family = "KoPubWorldDotum Medium")
pies_grid <- grid.arrange(pie_bureaus,
pie_commons,
ncol = 2,
top = "")
pies_grid## TableGrob (2 x 2) "arrange": 3 grobs
## z cells name grob
## 1 1 (2-2,1-1) arrange gtable[layout]
## 2 2 (2-2,2-2) arrange gtable[layout]
## 3 3 (1-1,1-2) arrange text[GRID.text.1273]ggsave("../pics/sejong_bureaus_commons_pie_ggplot.png", pies_grid, width = 8, height = 4, dpi = 72)
지역별 찬반
관료와 품관촌민
vote_region_bureaus %>%
kable(caption = "관료들의 지역별 찬반")| 서울 | 경기 | 평안 | 황해 | 충청 | 강원 | 함길 | 경상 | 전라 | |
|---|---|---|---|---|---|---|---|---|---|
| 찬성 | 723 | 29 | 6 | 17 | 35 | 5 | 3 | 55 | 42 |
| 반대 | 704 | 5 | 36 | 17 | 28 | 10 | 15 | 16 | 14 |
# kable(vote_region_bureaus, caption = "관료들의 지역별 찬반")
vote_region_bureaus %>%
prop.table(margin = 2) %>%
`*`(100) %>%
format(digits = 3, nsmall = 1) %>%
kable(align = "r", caption = "관료들의 지역별 찬반(%)")| 서울 | 경기 | 평안 | 황해 | 충청 | 강원 | 함길 | 경상 | 전라 | |
|---|---|---|---|---|---|---|---|---|---|
| 찬성 | 50.7 | 85.3 | 14.3 | 50.0 | 55.6 | 33.3 | 16.7 | 77.5 | 75.0 |
| 반대 | 49.3 | 14.7 | 85.7 | 50.0 | 44.4 | 66.7 | 83.3 | 22.5 | 25.0 |
# kable(format(prop.table(vote_region_bureaus, margin = 2) * 100, digits = 3, nsmall = 1),
# align = rep("r", 9), caption = "관료들의 지역별 찬반(%)")
vote_region_commons %>%
format(big.mark = ",") %>%
kable(align = "r", caption = "품관촌민들의 지역별 찬반")| 유후사 | 경기 | 평안 | 황해 | 충청 | 강원 | 함길 | 경상 | 전라 | |
|---|---|---|---|---|---|---|---|---|---|
| 찬성 | 1,123 | 17,076 | 1,326 | 4,454 | 6,982 | 939 | 75 | 36,262 | 29,505 |
| 반대 | 71 | 236 | 28,474 | 15,601 | 14,013 | 6,888 | 7,387 | 377 | 257 |
# kable(format(vote_region_commons, big.mark = ","),
# align = "r", caption = "품관촌민들의 지역별 찬반")
vote_region_commons %>%
prop.table(margin = 2) %>%
`*`(100) %>%
format(digits = 1, nsmall = 1) %>%
kable(align = "r", caption = "품관촌민들의 지역별 찬반(%)")| 유후사 | 경기 | 평안 | 황해 | 충청 | 강원 | 함길 | 경상 | 전라 | |
|---|---|---|---|---|---|---|---|---|---|
| 찬성 | 94.1 | 98.6 | 4.4 | 22.2 | 33.3 | 12.0 | 1.0 | 99.0 | 99.1 |
| 반대 | 5.9 | 1.4 | 95.6 | 77.8 | 66.7 | 88.0 | 99.0 | 1.0 | 0.9 |
# kable(format(prop.table(vote_region_commons, margin = 2) * 100, digits = 1, nsmall = 1),
# align = "r", caption = "품관촌민들의 지역별 찬반(%)")서울의 찬반
vote_seoul_class %>%
kable(caption = "서울의 찬반")| 대신 등 | 3품이하현직 | 3품이하전직 | |
|---|---|---|---|
| 찬성 | 21 | 259 | 443 |
| 반대 | 194 | 393 | 117 |
# kable(vote_seoul_class, caption = "서울의 찬반")
vote_seoul_class %>%
prop.table(margin = 2) %>%
`*`(100) %>%
format(digits = 1, nsmall = 1) %>%
kable(align = "r", caption = "서울의 찬반(%)")| 대신 등 | 3품이하현직 | 3품이하전직 | |
|---|---|---|---|
| 찬성 | 9.8 | 39.7 | 79.1 |
| 반대 | 90.2 | 60.3 | 20.9 |
# kable(format(prop.table(vote_seoul_class, margin = 2) * 100, digits = 1, nsmall = 1),
# align = "r", caption = "서울의 찬반(%)")
막대그래프
Stack
vote_seoul_df <- as.data.frame(vote_seoul_class)
n_fill <- length(levels(vote_seoul_df[, 1]))
x_stack <- vote_seoul_df[, 2]
y_stack <- unlist(tapply(vote_seoul_df[, 3],
vote_seoul_df[, 2],
function(x){x / 2 +
c(0, cumsum(head(x, -1)))}))
y_breaks <- unlist(tapply(vote_seoul_df[, 3],
vote_seoul_df[, 2],
cumsum))
y_label <- format(y_breaks, big.mark = ",")
b1_seoul <- ggplot(vote_seoul_df,
aes(x = x_stack,
y = vote_seoul_df[, 3],
fill = vote_seoul_df[, 1])) +
geom_bar(stat = "identity",
position = position_stack(reverse = TRUE))
b1_seoul
b2_seoul <- b1_seoul +
theme_bw(base_family = "KoPubWorldDotum Medium") +
# theme_kr +
scale_x_discrete(name = "계급") +
scale_y_continuous(name = "집계",
breaks = y_breaks,
labels = y_label) +
scale_fill_manual(name = "찬반",
values = rainbow(n_fill)[n_fill:1],
labels = vote_seoul_df[, 1])
b2_seoul 
b3_seoul <- b2_seoul +
# geom_text(aes(y = y_stack/2), label = vote_seoul_df$Freq)
geom_text(aes(y = y_stack),
label = vote_seoul_df[, 3],
position = "identity")
b3_seoul
ggsave("../pics/sejong_seoul_barplot_stack_ggplot.png", b3_seoul, width = 6, height = 4.5, dpi = 72)
Dodge
n_fill <- length(levels(vote_seoul_df[, 1]))
b1_seoul_dodge <- ggplot(vote_seoul_df,
aes(x = x_stack,
y = vote_seoul_df[, 3],
fill = vote_seoul_df[, 1])) +
geom_bar(stat = "identity", position = "dodge")
b1_seoul_dodge
b2_seoul_dodge <- b1_seoul_dodge +
theme_bw(base_family = "KoPubWorldDotum Medium") +
# theme_kr +
scale_x_discrete(name = "계급") +
scale_y_continuous(name = "집계",
breaks = vote_seoul_df[, 3],
labels = vote_seoul_df[, 3]) +
scale_fill_manual(name = "찬반",
values = rainbow(n_fill)[n_fill:1],
labels = vote_seoul_df[, 1])
b2_seoul_dodge 
# N <- nrow(vote_seoul_df)
# index <- as.vector(matrix(1:N, nrow = 2)[2:1, ])
y_label <- unlist(tapply(vote_seoul_df[, 3],
vote_seoul_df[, 2],
rev))
b3_seoul_dodge <- b2_seoul_dodge +
geom_text(aes(y = vote_seoul_df[, 3] / 2),
# label = vote_seoul_df[index, "Freq"],
label = y_label,
position = position_dodge(width = 0.9)) +
ggtitle("서울의 찬반")
b3_seoul_dodge
# ggsave("../pics/sejong_seoul_barplot_dodge_ggplot.png", b3_seoul_dodge, width = 6, height = 4.5, dpi = 72)
Fill
n_fill <- length(levels(vote_seoul_df[, 1]))
y_fill <- unlist(tapply(vote_seoul_df[, 3],
x_stack,
function(x){cumsum(x) / sum(x)}))
p_fill <- unlist(tapply(vote_seoul_df[, 3],
x_stack,
function(x){(x / 2 + c(0, cumsum(head(x, -1))))/sum(x)}))
b1_seoul_fill <- ggplot(vote_seoul_df,
aes(x = x_stack,
y = vote_seoul_df[, 3],
fill = vote_seoul_df[, 1])) +
geom_bar(stat = "identity",
position = position_fill(reverse = TRUE))
b1_seoul_fill
b2_seoul_fill <- b1_seoul_fill +
theme_bw(base_family = "KoPubWorldDotum Medium") +
# theme_kr +
scale_x_discrete(name = "계급") +
scale_y_continuous(name = "집계(%)",
breaks = y_fill,
labels = format(y_fill * 100,
digits = 2,
nsmall = 1)) +
scale_fill_manual(name = "찬반",
values = rainbow(n_fill)[n_fill:1],
labels = vote_seoul_df[, 1])
b2_seoul_fill 
b3_seoul_fill <- b2_seoul_fill +
geom_text(aes(y = p_fill),
label = vote_seoul_df[, 3],
position = "identity") +
ggtitle("서울의 찬반")
b3_seoul_fill
# ggsave("../pics/sejong_seoul_barplot_fill_ggplot.png", b3_seoul_fill, width = 6, height = 4.5, dpi = 72)
barplot_gg
Stack
barplot_gg_stack <- function(df,
base_family = "KoPubWorldDotum Medium",
ggtitle = "",
xlab = "",
ylab = "",
fill_name = ""){
n_fill <- length(levels(df[, 1]))
x <- df[, 2]
y <- unlist(tapply(df[, 3],
x,
function(x){x / 2 + c(0, cumsum(head(x, -1)))}))
y_breaks <- unlist(tapply(df[, 3],
x,
cumsum))
y_label <- format(y_breaks, big.mark = ",")
delta <- (max(y_breaks) - min(y_breaks)) / 20
y_breaks_sort <- sort(y_breaks)
diff(y_breaks_sort) < delta
index <- which(diff(y_breaks_sort) > delta)
y_breaks <- c(0, y_breaks_sort[c(index, length(y_breaks_sort))])
y_label <- format(y_breaks, big.mark = ",")
b1 <- ggplot(df,
aes(x = x,
y = df[, 3],
fill = df[, 1])) +
geom_bar(stat = "identity",
position = position_stack(reverse = TRUE))
b2 <- b1 +
theme_bw(base_family = base_family) +
# theme_kr +
scale_x_discrete(name = xlab) +
scale_y_continuous(name = ylab,
breaks = y_breaks,
labels = y_label) +
scale_fill_manual(name = fill_name,
values = rainbow(n_fill)[n_fill:1],
labels = df[, 1],
guide = guide_legend())
b3 <- b2 +
geom_text(aes(y = y),
label = format(ifelse(df[, 3] == 0, "", df[, 3]),
big.mark = ","),
position = "identity") +
ggtitle(ggtitle)
return(b3)
}
Dodge
barplot_gg_dodge <- function(df,
base_family = "KoPubWorldDotum Medium",
ggtitle = "",
xlab = "",
ylab = "",
fill_name = ""){
n_fill <- length(levels(df[, 1]))
x <- df[, 2]
y_dodge <- df[, 3]
b1 <- ggplot(df,
aes(x = x,
y = df[, 3],
fill = df[, 1])) +
geom_bar(stat = "identity",
position = "dodge")
b2 <- b1 +
theme_bw(base_family = base_family) +
# theme_kr +
scale_x_discrete(name = xlab) +
scale_y_continuous(name = ylab,
breaks = y_dodge,
labels = format(y_dodge, big.mark = ",")) +
scale_fill_manual(name = fill_name,
values = rainbow(n_fill)[n_fill:1],
labels = df[, 1])
# N <- nrow(df)
# index <- as.vector(matrix(1:N, nrow = 2)[2:1, ])
y_label <- unlist(tapply(df[, 3],
df[, 2],
rev))
b3 <- b2 +
geom_text(aes(y = y_dodge / 2),
# label = format(df[index, "Freq"], big.mark = ","),
label = ifelse(y_label == 0, "", y_label),
position = position_dodge(width = 0.9)) +
ggtitle(ggtitle)
return(b3)
}
Fill
barplot_gg_fill <- function(df,
base_family = "KoPubWorldDotum Medium",
ggtitle = "",
xlab = "",
ylab = "",
fill_name = ""){
n_fill <- length(levels(df[, 1]))
x <- df[, 2]
y_fill <- unlist(tapply(df[, 3],
x,
function(x){cumsum(x) / sum(x)}))
p_fill <- unlist(tapply(df[, 3],
x,
function(x){(x / 2 + c(0, cumsum(head(x, -1))))/sum(x)}))
b1 <- ggplot(df,
aes(x = x,
y = df[, 3],
fill = df[, 1])) +
geom_bar(stat = "identity",
position = position_fill(reverse = TRUE))
b2 <- b1 +
theme_bw(base_family = base_family) +
# theme_kr +
scale_x_discrete(name = xlab) +
scale_y_continuous(name = ylab,
breaks = y_fill,
labels = format(y_fill * 100,
digits = 2,
nsmall = 1)) +
scale_fill_manual(name = fill_name,
values = rainbow(n_fill)[n_fill:1],
labels = df[, 1],
guide = guide_legend())
b3 <- b2 +
geom_text(aes(y = p_fill),
label = format(ifelse(df[, 3] == 0, "", df[, 3]),
big.mark = ","),
position = "identity") +
ggtitle(ggtitle)
return(b3)
}
barplot_gg
barplot_gg <- function(x,
position,
base_family = "KoPubWorldDotum Medium",
ggtitle = "",
xlab = "",
ylab = "",
fill_name = ""){
switch(position,
stack = barplot_gg_stack(x,
base_family = base_family,
ggtitle = ggtitle,
xlab = xlab,
ylab = ylab,
fill_name = fill_name),
dodge = barplot_gg_dodge(x,
base_family = base_family,
ggtitle = ggtitle,
xlab = xlab,
ylab = ylab,
fill_name = fill_name),
fill = barplot_gg_fill(x,
base_family = base_family,
ggtitle = ggtitle,
xlab = xlab,
ylab = ylab,
fill_name = fill_name))
}
(bar_seoul_stack <- barplot_gg(vote_seoul_df,
position = "stack",
base_family = "KoPubWorldDotum Medium",
ggtitle = "서울의 찬반",
xlab = "계급",
ylab = "집계",
fill_name = "찬반"))
# ggsave("../pics/sejong_seoul_barplot_stack_ggplotv2.png", bar_seoul_stack, width = 6, height = 4.5, dpi = 72)
(bar_seoul_dodge <- barplot_gg(vote_seoul_df,
position = "dodge",
base_family = "KoPubWorldDotum Medium",
ggtitle = "서울의 찬반",
xlab = "계급",
ylab = "집계",
fill_name = "찬반"))
# ggsave("../pics/sejong_seoul_barplot_dodge_ggplotv2.png", bar_seoul_dodge, width = 6, height = 4.5, dpi = 72)
(bar_seoul_fill <- barplot_gg(vote_seoul_df,
position = "fill",
base_family = "KoPubWorldDotum Medium",
ggtitle = "서울의 찬반",
xlab = "계급",
ylab = "집계(%)",
fill_name = "찬반"))
# ggsave("../pics/sejong_seoul_barplot_fill_ggplotv2.png", bar_seoul_fill, width = 6, height = 4.5, dpi = 72)
dump(c("barplot_gg","barplot_gg_stack", "barplot_gg_dodge", "barplot_gg_fill"),
file = "./barplot_gg.R")
mosaic plot
Data
vote_seoul_class_df <- vote_seoul_class %>%
as.data.frame
vote_seoul_class_p_df <- vote_seoul_class %>%
prop.table %>%
as.data.frame
vote_seoul_class_p_m <- vote_seoul_class_df %>%
`[`(, 3) %>%
tapply(vote_seoul_class_df[, 2], sum) %>%
prop.table
# vote_seoul_class_df <- as.data.frame(vote_seoul_class)
# vote_seoul_class_sum <- tapply(vote_seoul_class_df[, 3],
# vote_seoul_class_df[, 2],
# sum)
# vote_seoul_class_p_m <- prop.table(vote_seoul_class_sum)
# vote_seoul_class_p <- prop.table(vote_seoul_class)
# vote_seoul_class_p_2 <- prop.table(vote_seoul_class,
# margin = 2)
# vote_seoul_class_p_df <- as.data.frame(vote_seoul_class_p)
vote_seoul_class_p_df$width <- vote_seoul_class_p_m[match(vote_seoul_class_p_df[, 2],
names(vote_seoul_class_p_m))]
# vote_seoul_class_p_2_df <- as.data.frame(vote_seoul_class_p_2)
vote_seoul_class_p_df$height <- vote_seoul_class %>%
prop.table(margin = 2) %>%
as.data.frame %>%
`[`(, 3)
# vote_seoul_class_p_df$height <- vote_seoul_class_p_2_df[, 3]
vote_seoul_class_p_df$label_height <- unlist(tapply(vote_seoul_class_p_df$height, vote_seoul_class_p_df[, 2], function(x) x / 2 + c(0, cumsum(head(x, -1)))))
vote_seoul_class_p_df$y_breaks <- unlist(tapply(vote_seoul_class_p_df$height,
vote_seoul_class_p_df[, 2],
cumsum))
x_center <- vote_seoul_class_p_m / 2 + c(0, cumsum(head(vote_seoul_class_p_m, -1)))
# x_center <- (cumsum(vote_seoul_class_p_m) +
# c(0, head(cumsum(vote_seoul_class_p_m), -1)))/2
vote_seoul_class_p_df$center <- x_center[match(vote_seoul_class_p_df[, 2],
names(x_center))]
m1 <- ggplot(vote_seoul_class_p_df,
aes(x = center,
y = height,
width = width)) +
geom_bar(aes(fill = vote_seoul_class_p_df[, 1]),
stat = "identity",
col = "white",
size = 2,
position = position_stack(reverse = TRUE))
m1
m2 <- m1 +
theme_bw(base_family = "KoPubWorldDotum Medium")
m2
m3 <- m2 +
geom_text(aes(x = center,
y = 1.05),
label = vote_seoul_class_p_df[, 2],
family = "KoPubWorldDotum Medium")
m3
m4 <- m3 +
geom_text(aes(x = center,
y = label_height),
label = format(vote_seoul_class_df[, 3],
big.mark = ","),
position = position_identity())
m4
x_breaks <- c(0, ifelse(cumsum(vote_seoul_class_p_m) < 0.1, 0.0, cumsum(vote_seoul_class_p_m)))
x_label <- format(x_breaks * 100,
digits = 3,
nsmall = 1)
y_breaks <- vote_seoul_class_p_df$y_breaks
delta <- (max(y_breaks) - min(y_breaks)) / 20
y_breaks_sort <- sort(y_breaks)
index <- which(diff(y_breaks_sort) > delta)
y_breaks <- c(0, y_breaks_sort[c(index, length(y_breaks_sort))])
y_label <- format(y_breaks * 100,
digits = 2,
nsmall = 1)
m5 <- m4 +
scale_x_continuous(name = "계급(누적 %)",
breaks = x_breaks,
label = x_label) +
scale_y_continuous(name = "찬반(%)",
breaks = y_breaks,
label = y_label) +
scale_fill_manual(name = "찬반",
values = rainbow(2)[2:1],
labels = vote_seoul_df[, 1],
guide = guide_legend()) +
ggtitle("서울의 찬반")
m5
mosaic_gg
mosaic_gg <- function(tbl, base_family = "KoPubWorldDotum Medium",
ggtitle = "",
xlab = "",
ylab = "",
fill_name = ""){
tbl_df <- tbl %>%
as.data.frame
N <- length(levels(tbl_df[, 1]))
tbl_p_df <- tbl %>%
prop.table %>%
as.data.frame
tbl_p_m <- tbl_df %>%
`[`(, 3) %>%
tapply(tbl_df[, 2], sum) %>%
prop.table
tbl_p_df$width <- tbl_p_m[match(tbl_p_df[, 2], names(tbl_p_m))]
tbl_p_df$height <- tbl %>%
prop.table(margin = 2) %>%
as.data.frame %>%
`[`(, 3)
tbl_p_df$label_height <- unlist(tapply(tbl_p_df$height,
tbl_p_df[, 2],
function(x) x / 2 + c(0, cumsum(head(x, -1)))))
tbl_p_df$y_breaks <- unlist(tapply(tbl_p_df$height,
tbl_p_df[, 2],
cumsum))
x_center <- tbl_p_m / 2 + c(0, cumsum(head(tbl_p_m, -1)))
# x_center <- (cumsum(tbl_p_m) + c(0, head(cumsum(tbl_p_m), -1)))/2
tbl_p_df$center <- x_center[match(tbl_p_df[, 2], names(x_center))]
m1 <- ggplot(tbl_p_df,
aes(x = center, y = height, width = width)) +
geom_bar(aes(fill = tbl_df[, 1]),
stat = "identity",
col = "white",
size = 1,
position = position_stack(reverse = TRUE))
m2 <- m1 +
theme_bw(base_family = base_family)
m3 <- m2 +
geom_text(aes(x = center, y = 1.05),
label = ifelse(tbl_p_df[, 2] == 0, "", tbl_p_df[, 2]),
family = base_family)
m4 <- m3 +
geom_text(aes(x = center, y = label_height),
label = format(ifelse(tbl_df[, 3] == 0, "", tbl_df[, 3]),
big.mark = ","),
position = position_identity())
x_breaks <- c(0, ifelse(cumsum(tbl_p_m) < 0.1, 0.0, cumsum(tbl_p_m)))
x_label <- format(x_breaks * 100,
digits = 3,
nsmall = 1)
y_breaks <- tbl_p_df$y_breaks
delta <- (max(y_breaks) - min(y_breaks)) / 20
y_breaks_sort <- sort(y_breaks)
diff(y_breaks_sort) < delta
index <- which(diff(y_breaks_sort) > delta)
y_breaks <- c(0, y_breaks_sort[c(index, length(y_breaks_sort))])
y_label <- format(y_breaks * 100,
digits = 2,
nsmall = 1)
m5 <- m4 +
scale_x_continuous(name = xlab,
breaks = x_breaks,
label = x_label) +
scale_y_continuous(name = ylab,
breaks = y_breaks,
label = y_label) +
scale_fill_manual(name = fill_name,
values = rainbow(N)[N:1],
labels = tbl_df[, 1],
guide = guide_legend()) +
ggtitle(ggtitle) +
theme(plot.margin = unit(c(1, 2, 1, 1), "lines"))
return(m5)
}
dump(list = "mosaic_gg", file = "./mosaic_gg.R")source("./mosaic_gg.R")
m5_seoul <- mosaic_gg(vote_seoul_class,
base_family = "KoPubWorldDotum Medium",
ggtitle = "서울의 찬반",
xlab = "계급",
ylab = "찬반",
fill_name = "찬반")
m5_seoul
ggsave("../pics/sejong_seoul_mosaic_ggplot.png", m5_seoul, width = 6, height = 4.5, dpi = 72)
지역별 관료들의 찬반(서울 제외)
vote_region_bureaus %>%
`[`(, -1) %>%
kable(caption = "지역별 관료들의 찬반(서울 제외)")| 경기 | 평안 | 황해 | 충청 | 강원 | 함길 | 경상 | 전라 | |
|---|---|---|---|---|---|---|---|---|
| 찬성 | 29 | 6 | 17 | 35 | 5 | 3 | 55 | 42 |
| 반대 | 5 | 36 | 17 | 28 | 10 | 15 | 16 | 14 |
# kable(vote_region_bureaus[, -1],
# caption = "지역별 관료들의 찬반(서울 제외)")
vote_region_bureaus %>%
`[`(, -1) %>%
prop.table(margin = 2) %>%
`*`(100) %>%
format(digits = 1, nsmall = 1) %>%
kable(align = "r", caption = "지역별 관료들의 찬반(서울 제외, %)")| 경기 | 평안 | 황해 | 충청 | 강원 | 함길 | 경상 | 전라 | |
|---|---|---|---|---|---|---|---|---|
| 찬성 | 85.3 | 14.3 | 50.0 | 55.6 | 33.3 | 16.7 | 77.5 | 75.0 |
| 반대 | 14.7 | 85.7 | 50.0 | 44.4 | 66.7 | 83.3 | 22.5 | 25.0 |
# kable(format(prop.table(vote_region_bureaus[, -1],
# margin = 2)*100,
# digits = 1,
# nsmall = 1),
# align = "r",
# caption = "지역별 관료들의 찬반(서울 제외, %)")
막대그래프
vote_region_bureaus.df <- as.data.frame(vote_region_bureaus[, -1])
barplot_gg(vote_region_bureaus.df,
position = "stack",
base_family = "KoPubWorldDotum Medium",
ggtitle = "지방 관료들의 찬반",
xlab = "지역",
ylab = "집계",
fill_name = "찬반")
ggsave("../pics/sejong_bureaus_barplot_stack_ggplot.png", width = 9, height = 4.5, dpi = 72)
mosaic plot
m5_bureaus <- mosaic_gg(vote_region_bureaus[, -1],
base_family = "KoPubWorldDotum Medium",
ggtitle = "지방관료의 찬반",
xlab = "지역별 관료들의 비중(누적 %)",
ylab = "찬반(%)",
fill_name = "찬반")
m5_bureaus
ggsave("../pics/sejong_bureaus_mosaic_ggplot.png", m5_bureaus, width = 9, height = 4.5, dpi = 72)
품관촌민들의 지역별 찬반
막대그래프
vote_region_commons_df <- as.data.frame(vote_region_commons)
barplot_gg(vote_region_commons_df,
position = "stack",
base_family = "KoPubWorldDotum Medium",
ggtitle = "품관촌민의 찬반",
xlab = "지역",
ylab = "집계",
fill_name = "찬반")
ggsave("../pics/sejong_commons_barplot_stack_ggplot.png", width = 9, height = 4.5, dpi = 72)
mosaic plot
m5_commons <- mosaic_gg(vote_region_commons,
base_family = "KoPubWorldDotum Medium",
ggtitle = "품관촌민의 찬반",
xlab = "지역별 품관촌민의 비중(누적%)",
ylab = "찬반(%)",
fill_name = "찬반")
m5_commons
ggsave("../pics/sejong_commons_mosaic_ggplot.png", m5_commons, width = 12, height = 4.5, dpi = 72)
충청도
관료들의 찬반과 품관촌민의 찬반이 다른 곳.
vote_chung_class %>%
format(big.mark = ",") %>%
kable(align = "r", caption = "충청도 사례")| 대신 등 | 수령 | 품관촌민 | |
|---|---|---|---|
| 찬성 | 0 | 35 | 6,982 |
| 반대 | 2 | 26 | 14,013 |
# kable(format(vote_chung_class,
# big.mark = ","),
# caption = "충청도 사례",
# align = "r")
vote_chung_class %>%
prop.table(margin = 2) %>%
`*`(100) %>%
format(digits = 3, nsmall = 1) %>%
kable(align = "r", caption = "충청도 사례")| 대신 등 | 수령 | 품관촌민 | |
|---|---|---|---|
| 찬성 | 0.0 | 57.4 | 33.3 |
| 반대 | 100.0 | 42.6 | 66.7 |
# kable(format(prop.table(vote_chung_class,
# margin = 2)*100,
# digits = 3, nsmall = 1),
# caption = "충청도 사례",
# align = "r")막대그래프
vote_chung_class_df <- as.data.frame(vote_chung_class)
barplot_gg(vote_chung_class_df,
position = "fill",
base_family = "KoPubWorldDotum Medium",
ggtitle = "충청도의 찬반",
xlab = "계급",
ylab = "집계(%)",
fill_name = "찬반")
ggsave("../pics/sejong_chungcheong_barplot_fill_ggplot.png", width = 6, height = 4.5, dpi = 72)
mosaic plot
m5_chungcheong <- mosaic_gg(vote_chung_class[, -1],
base_family = "KoPubWorldDotum Medium",
ggtitle = "충청도의 찬반",
xlab = "계급(누적%)",
ylab = "찬반(%)",
fill_name = "찬반")
m5_chungcheong
ggsave("../pics/sejong_chungcheong_mosaic_ggplot.png", m5_chungcheong, width = 12, height = 4.5, dpi = 72)자료 정리
save.image(file = "./sejong_ggplot.RData")