Gini_OECD

Data 작업

OECD 국가들의 Gini계수 읽어들이기. 세전과 세후로 구분. 자료구조로 인하여 sep="\t"을 사용한 것에 유의

Gini.b.tax<-read.table(file="../data/Gini_before_tax.txt", header=F, sep="\t")
Gini.a.tax<-read.table(file="../data/Gini_after_tax.txt", header=F, sep="\t")
str(Gini.b.tax)

## 'data.frame':    34 obs. of  8 variables:
##  $ V1: chr  "Australia" "Austria" "Belgium" "Canada" ...
##  $ V2: num  NA NA NA 0.385 NA NA NA NA 0.343 NA ...
##  $ V3: num  NA NA 0.449 0.395 NA NA 0.373 NA 0.387 0.38 ...
##  $ V4: num  NA NA NA 0.403 NA NA 0.396 NA NA 0.37 ...
##  $ V5: num  0.467 NA 0.472 0.43 0.441 0.442 0.417 NA 0.479 0.473 ...
##  $ V6: num  0.476 NA 0.464 0.44 NA 0.472 0.415 NA 0.478 0.49 ...
##  $ V7: num  0.465 0.433 0.494 0.436 0.414 0.474 0.417 0.504 0.483 0.485 ...
##  $ V8: num  0.468 0.472 0.469 0.441 0.426 0.444 0.416 0.458 0.465 0.483 ...

str(Gini.a.tax)

## 'data.frame':    34 obs. of  8 variables:
##  $ V1: chr  "Australia" "Austria" "Belgium" "Canada" ...
##  $ V2: num  NA NA NA 0.304 NA NA NA NA 0.235 NA ...
##  $ V3: num  NA 0.236 0.274 0.293 NA NA 0.221 NA 0.209 0.3 ...
##  $ V4: num  NA NA NA 0.287 NA 0.232 0.226 NA NA 0.29 ...
##  $ V5: num  0.309 0.238 0.287 0.289 0.427 0.257 0.215 NA 0.218 0.277 ...
##  $ V6: num  0.317 0.252 0.289 0.318 NA 0.26 0.226 NA 0.247 0.287 ...
##  $ V7: num  0.315 0.265 0.271 0.317 0.403 0.268 0.232 0.349 0.254 0.288 ...
##  $ V8: num  0.336 0.261 0.259 0.324 0.394 0.256 0.248 0.315 0.259 0.293 ...

2000년 후반 자료만 모아서 새로운 data frame 구성

(Gini.b.a<-data.frame(Country=Gini.b.tax$V1, Before=Gini.b.tax$V8, After=Gini.a.tax$V8))

##            Country Before After
## 1        Australia  0.468 0.336
## 2          Austria  0.472 0.261
## 3          Belgium  0.469 0.259
## 4           Canada  0.441 0.324
## 5            Chile  0.426 0.394
## 6   Czech_Republic  0.444 0.256
## 7          Denmark  0.416 0.248
## 8          Estonia  0.458 0.315
## 9          Finland  0.465 0.259
## 10          France  0.483 0.293
## 11         Germany  0.504 0.295
## 12          Greece  0.436 0.307
## 13         Hungary  0.466 0.272
## 14         Iceland  0.382 0.301
## 15         Ireland     NA 0.293
## 16          Israel  0.498 0.371
## 17           Italy  0.534 0.337
## 18           Japan  0.462 0.329
## 19      Luxembourg  0.482 0.288
## 20          Mexico  0.494 0.476
## 21     Netherlands  0.426 0.294
## 22     New_Zealand  0.455 0.330
## 23          Norway  0.410 0.250
## 24          Poland  0.470 0.305
## 25        Portugal  0.521 0.353
## 26 Slovak_Republic  0.416 0.257
## 27        Slovenia  0.423 0.236
## 28     South_Korea  0.344 0.315
## 29           Spain  0.461 0.317
## 30          Sweden  0.426 0.259
## 31     Switzerland  0.409 0.303
## 32          Turkey  0.470 0.409
## 33  United_Kingdom  0.456 0.345
## 34   United_States  0.486 0.378

세전과 세후의 Gini 계수 차이를 개선도(Improvement)라고 명명.

Gini.b.a$Improvement<-Gini.b.a[,2]-Gini.b.a[,3]
Gini.b.a

##            Country Before After Improvement
## 1        Australia  0.468 0.336       0.132
## 2          Austria  0.472 0.261       0.211
## 3          Belgium  0.469 0.259       0.210
## 4           Canada  0.441 0.324       0.117
## 5            Chile  0.426 0.394       0.032
## 6   Czech_Republic  0.444 0.256       0.188
## 7          Denmark  0.416 0.248       0.168
## 8          Estonia  0.458 0.315       0.143
## 9          Finland  0.465 0.259       0.206
## 10          France  0.483 0.293       0.190
## 11         Germany  0.504 0.295       0.209
## 12          Greece  0.436 0.307       0.129
## 13         Hungary  0.466 0.272       0.194
## 14         Iceland  0.382 0.301       0.081
## 15         Ireland     NA 0.293          NA
## 16          Israel  0.498 0.371       0.127
## 17           Italy  0.534 0.337       0.197
## 18           Japan  0.462 0.329       0.133
## 19      Luxembourg  0.482 0.288       0.194
## 20          Mexico  0.494 0.476       0.018
## 21     Netherlands  0.426 0.294       0.132
## 22     New_Zealand  0.455 0.330       0.125
## 23          Norway  0.410 0.250       0.160
## 24          Poland  0.470 0.305       0.165
## 25        Portugal  0.521 0.353       0.168
## 26 Slovak_Republic  0.416 0.257       0.159
## 27        Slovenia  0.423 0.236       0.187
## 28     South_Korea  0.344 0.315       0.029
## 29           Spain  0.461 0.317       0.144
## 30          Sweden  0.426 0.259       0.167
## 31     Switzerland  0.409 0.303       0.106
## 32          Turkey  0.470 0.409       0.061
## 33  United_Kingdom  0.456 0.345       0.111
## 34   United_States  0.486 0.378       0.108

개선도가 낮은 순서로 나열. 아일랜드는 세전 자료가 없기 때문에 맨 뒤로 위치.

Gini.b.a[order(Gini.b.a$Improvement), ]

##            Country Before After Improvement
## 20          Mexico  0.494 0.476       0.018
## 28     South_Korea  0.344 0.315       0.029
## 5            Chile  0.426 0.394       0.032
## 32          Turkey  0.470 0.409       0.061
## 14         Iceland  0.382 0.301       0.081
## 31     Switzerland  0.409 0.303       0.106
## 34   United_States  0.486 0.378       0.108
## 33  United_Kingdom  0.456 0.345       0.111
## 4           Canada  0.441 0.324       0.117
## 22     New_Zealand  0.455 0.330       0.125
## 16          Israel  0.498 0.371       0.127
## 12          Greece  0.436 0.307       0.129
## 1        Australia  0.468 0.336       0.132
## 21     Netherlands  0.426 0.294       0.132
## 18           Japan  0.462 0.329       0.133
## 8          Estonia  0.458 0.315       0.143
## 29           Spain  0.461 0.317       0.144
## 26 Slovak_Republic  0.416 0.257       0.159
## 23          Norway  0.410 0.250       0.160
## 24          Poland  0.470 0.305       0.165
## 30          Sweden  0.426 0.259       0.167
## 7          Denmark  0.416 0.248       0.168
## 25        Portugal  0.521 0.353       0.168
## 27        Slovenia  0.423 0.236       0.187
## 6   Czech_Republic  0.444 0.256       0.188
## 10          France  0.483 0.293       0.190
## 13         Hungary  0.466 0.272       0.194
## 19      Luxembourg  0.482 0.288       0.194
## 17           Italy  0.534 0.337       0.197
## 9          Finland  0.465 0.259       0.206
## 11         Germany  0.504 0.295       0.209
## 3          Belgium  0.469 0.259       0.210
## 2          Austria  0.472 0.261       0.211
## 15         Ireland     NA 0.293          NA

개선도가 높은 순서로 나라명을 나열하려면, decreasing = TRUE 추가.

Gini.b.a[order(Gini.b.a$Improvement, decreasing=TRUE), ]

##            Country Before After Improvement
## 2          Austria  0.472 0.261       0.211
## 3          Belgium  0.469 0.259       0.210
## 11         Germany  0.504 0.295       0.209
## 9          Finland  0.465 0.259       0.206
## 17           Italy  0.534 0.337       0.197
## 13         Hungary  0.466 0.272       0.194
## 19      Luxembourg  0.482 0.288       0.194
## 10          France  0.483 0.293       0.190
## 6   Czech_Republic  0.444 0.256       0.188
## 27        Slovenia  0.423 0.236       0.187
## 25        Portugal  0.521 0.353       0.168
## 7          Denmark  0.416 0.248       0.168
## 30          Sweden  0.426 0.259       0.167
## 24          Poland  0.470 0.305       0.165
## 23          Norway  0.410 0.250       0.160
## 26 Slovak_Republic  0.416 0.257       0.159
## 29           Spain  0.461 0.317       0.144
## 8          Estonia  0.458 0.315       0.143
## 18           Japan  0.462 0.329       0.133
## 1        Australia  0.468 0.336       0.132
## 21     Netherlands  0.426 0.294       0.132
## 12          Greece  0.436 0.307       0.129
## 16          Israel  0.498 0.371       0.127
## 22     New_Zealand  0.455 0.330       0.125
## 4           Canada  0.441 0.324       0.117
## 33  United_Kingdom  0.456 0.345       0.111
## 34   United_States  0.486 0.378       0.108
## 31     Switzerland  0.409 0.303       0.106
## 14         Iceland  0.382 0.301       0.081
## 32          Turkey  0.470 0.409       0.061
## 5            Chile  0.426 0.394       0.032
## 28     South_Korea  0.344 0.315       0.029
## 20          Mexico  0.494 0.476       0.018
## 15         Ireland     NA 0.293          NA

Graphic representation

세전 세후 Gini 계수를 시각적으로 비교하려면 barplot()이 적합함. barplot(height, ...)에서 height가 매트릭스일 때는 막대는 열의 각 요소를 크기대로 쌓아놓은 형태가 되므로, t()를 이용하여 transpose시킨 후 barplot()을 적용. 또한 transpose를 시켜도 여전히 data frame 이기 때문에 매트릭스로 강제 변환함. 세전, 세후 비교를 위해 쌓아 놓기 보다는 옆에 늘어세우는 게 나으므로 beside=TRUE를 적용하고 각 막대의 이름으로 나라이름을 사용.

barplot(as.matrix(t(Gini.b.a[, 2:3])), beside=TRUE, names.arg=Gini.b.a$Country)

개선도 순서(내림차순)를 o.improvement로 저장하여 지속적으로 활용.

o.improvement<-order(Gini.b.a$Improvement, decreasing=TRUE)
Gini.b.a$Country[o.improvement]

##  [1] "Austria"         "Belgium"         "Germany"        
##  [4] "Finland"         "Italy"           "Hungary"        
##  [7] "Luxembourg"      "France"          "Czech_Republic" 
## [10] "Slovenia"        "Portugal"        "Denmark"        
## [13] "Sweden"          "Poland"          "Norway"         
## [16] "Slovak_Republic" "Spain"           "Estonia"        
## [19] "Japan"           "Australia"       "Netherlands"    
## [22] "Greece"          "Israel"          "New_Zealand"    
## [25] "Canada"          "United_Kingdom"  "United_States"  
## [28] "Switzerland"     "Iceland"         "Turkey"         
## [31] "Chile"           "South_Korea"     "Mexico"         
## [34] "Ireland"

개선도 순서대로 막대를 늘어세우면,

barplot(as.matrix(t(Gini.b.a[o.improvement, 2:3])), beside=TRUE, names.arg=Gini.b.a$Country[o.improvement])

las=2를 이용하여 막대 이름을 눕힘.

barplot(as.matrix(t(Gini.b.a[o.improvement, 2:3])), beside=TRUE, names.arg=Gini.b.a$Country[o.improvement], las=2)

나라 이름이 가리지 않도록 par("mai")를 조정

old.par<-par(no.readonly=TRUE)
par("mai")

## [1] 1.02 0.82 0.82 0.42

par("mai"= c(1.5, 0.8, 0.8, 0.4))
barplot(as.matrix(t(Gini.b.a[o.improvement, 2:3])), beside=TRUE, names.arg=Gini.b.a$Country[o.improvement], las=2)

par(old.par)

불평등이 심하다고 판단하는 Gini 계수 0.4를 경계로 나눠 보면,

old.par<-par(no.readonly=TRUE)
par("mai")

## [1] 1.02 0.82 0.82 0.42

par("mai"= c(1.5, 0.8, 0.8, 0.4))
barplot(as.matrix(t(Gini.b.a[o.improvement, 2:3])), beside=TRUE, names.arg=Gini.b.a$Country[o.improvement], las=2)
abline(h=0.4, lty=2, col="red")

par(old.par)

범례와 메인 타이틀 추가. 좌표에 유의

old.par<-par(no.readonly=TRUE)
par("mai")

## [1] 1.02 0.82 0.82 0.42

par("mai"= c(1.5, 0.8, 0.8, 0.4))
barplot(as.matrix(t(Gini.b.a[o.improvement, 2:3])), beside=TRUE, names.arg=Gini.b.a$Country[o.improvement], legend.text=c("Before Tax", "After Tax"), args.legend=list(x=105, y=0.62), las=2)
abline(h=0.4, lty=2, col="red")
title(main="Gini Coefficients of OECD Countries")

par(old.par)

이번에는 막대를 눕히는 방법을 생각해 보자. 옆으로 눕히면서 las = 1 로 설정하면,

barplot(as.matrix(t(Gini.b.a[o.improvement, 2:3])), beside=TRUE, horiz=TRUE, names.arg=Gini.b.a$Country[o.improvement], las=1)

역시 나라 이름이 가리지 않도록 par("mai")를 조정.

old.par<-par(no.readonly=TRUE)
par("mai")

## [1] 1.02 0.82 0.82 0.42

par("mai"= c(1.0, 1.5, 0.8, 0.4))
barplot(as.matrix(t(Gini.b.a[o.improvement, 2:3])), beside=TRUE, horiz=TRUE, names.arg=Gini.b.a$Country[o.improvement], las=1)

par(old.par)

개선도가 낮은 순서대로 밑에서 올라가도록 다시 그리면,

old.par<-par(no.readonly=TRUE)
par("mai")

## [1] 1.02 0.82 0.82 0.42

par("mai"= c(1.0, 1.5, 0.8, 0.4))
barplot(as.matrix(t(Gini.b.a[order(Gini.b.a$Improvement, na.last=FALSE), 2:3])), beside=TRUE, horiz=TRUE, names.arg=Gini.b.a$Country[order(Gini.b.a$Improvement, na.last=FALSE)], las=1)

par(old.par)

이 때, Ireland가 맨 위에 올라오는 게 보기 좋지 않으므로, na.last=FALSE를 추가한 것임.
- 세전 Gini 계수 0.4를 경계로 나눠보면

old.par<-par(no.readonly=TRUE)
par("mai")

## [1] 1.02 0.82 0.82 0.42

par("mai"= c(1.0, 1.5, 0.8, 0.4))
barplot(as.matrix(t(Gini.b.a[order(Gini.b.a$Improvement, na.last=FALSE), 2:3])), beside=TRUE, horiz=TRUE, names.arg=Gini.b.a$Country[order(Gini.b.a$Improvement, na.last=FALSE)], las=1)
abline(v=0.4, lty=2, col="red")

par(old.par)

범례 및 메인 타이틀 추가. 시행착오를 거쳐 구한 좌표에 유의할 것.

old.par<-par(no.readonly=TRUE)
par("mai")

## [1] 1.02 0.82 0.82 0.42

par("mai"= c(1.0, 1.5, 0.8, 0.8))
barplot(as.matrix(t(Gini.b.a[order(Gini.b.a$Improvement, na.last=FALSE), 2:3])), beside=TRUE, horiz=TRUE, names.arg=Gini.b.a$Country[order(Gini.b.a$Improvement, na.last=FALSE)], legend.text=c("Before Tax", "After Tax"), args.legend=list(x=0.67, y=110), las=1)
abline(v=0.4, lty=2, col="red")
title(main="Gini Coefficients of OECD Countries")

par(old.par)

ggplot

Data reshaping

reshape2 package 를 검색 목록에 등록

library(reshape2)
(Gini.b.a.melt <- melt(Gini.b.a, id.vars = "Country", measure.vars = c("Before", "After"), variable.name = "Tax", value.name = "Gini.Coef"))

##            Country    Tax Gini.Coef
## 1        Australia Before     0.468
## 2          Austria Before     0.472
## 3          Belgium Before     0.469
## 4           Canada Before     0.441
## 5            Chile Before     0.426
## 6   Czech_Republic Before     0.444
## 7          Denmark Before     0.416
## 8          Estonia Before     0.458
## 9          Finland Before     0.465
## 10          France Before     0.483
## 11         Germany Before     0.504
## 12          Greece Before     0.436
## 13         Hungary Before     0.466
## 14         Iceland Before     0.382
## 15         Ireland Before        NA
## 16          Israel Before     0.498
## 17           Italy Before     0.534
## 18           Japan Before     0.462
## 19      Luxembourg Before     0.482
## 20          Mexico Before     0.494
## 21     Netherlands Before     0.426
## 22     New_Zealand Before     0.455
## 23          Norway Before     0.410
## 24          Poland Before     0.470
## 25        Portugal Before     0.521
## 26 Slovak_Republic Before     0.416
## 27        Slovenia Before     0.423
## 28     South_Korea Before     0.344
## 29           Spain Before     0.461
## 30          Sweden Before     0.426
## 31     Switzerland Before     0.409
## 32          Turkey Before     0.470
## 33  United_Kingdom Before     0.456
## 34   United_States Before     0.486
## 35       Australia  After     0.336
## 36         Austria  After     0.261
## 37         Belgium  After     0.259
## 38          Canada  After     0.324
## 39           Chile  After     0.394
## 40  Czech_Republic  After     0.256
## 41         Denmark  After     0.248
## 42         Estonia  After     0.315
## 43         Finland  After     0.259
## 44          France  After     0.293
## 45         Germany  After     0.295
## 46          Greece  After     0.307
## 47         Hungary  After     0.272
## 48         Iceland  After     0.301
## 49         Ireland  After     0.293
## 50          Israel  After     0.371
## 51           Italy  After     0.337
## 52           Japan  After     0.329
## 53      Luxembourg  After     0.288
## 54          Mexico  After     0.476
## 55     Netherlands  After     0.294
## 56     New_Zealand  After     0.330
## 57          Norway  After     0.250
## 58          Poland  After     0.305
## 59        Portugal  After     0.353
## 60 Slovak_Republic  After     0.257
## 61        Slovenia  After     0.236
## 62     South_Korea  After     0.315
## 63           Spain  After     0.317
## 64          Sweden  After     0.259
## 65     Switzerland  After     0.303
## 66          Turkey  After     0.409
## 67  United_Kingdom  After     0.345
## 68   United_States  After     0.378

str(Gini.b.a.melt)

## 'data.frame':    68 obs. of  3 variables:
##  $ Country  : chr  "Australia" "Austria" "Belgium" "Canada" ...
##  $ Tax      : Factor w/ 2 levels "Before","After": 1 1 1 1 1 1 1 1 1 1 ...
##  $ Gini.Coef: num  0.468 0.472 0.469 0.441 0.426 0.444 0.416 0.458 0.465 0.483 ...

ggplot2 등록 후 geom_bar()

library(ggplot2)
ggplot(Gini.b.a.melt, aes(x = Country, y = Gini.Coef, fill = Tax)) + geom_bar(stat = "identity", position = "dodge") +
  coord_flip()

개선도 순서대로 늘어세우려면 그 순서를 level로 갖는 factor로 만들어야함. o.improvement가 내림차순으로 정리되어 있는 순서이기 때문에 rev(o.improvement)는 올림차순으로 정리되어 있는 순서임. 따라서,

Gini.b.a$Country.order <- factor(Gini.b.a$Country, levels = Gini.b.a$Country[rev(o.improvement)])
Gini.b.a.order.melt <- melt(Gini.b.a, id.vars = "Country.order", measure.vars = c("Before", "After"), variable.name = "Tax", value.name = "Gini.Coef")
str(Gini.b.a.order.melt)

## 'data.frame':    68 obs. of  3 variables:
##  $ Country.order: Factor w/ 34 levels "Ireland","Mexico",..: 15 34 33 10 4 26 23 17 31 27 ...
##  $ Tax          : Factor w/ 2 levels "Before","After": 1 1 1 1 1 1 1 1 1 1 ...
##  $ Gini.Coef    : num  0.468 0.472 0.469 0.441 0.426 0.444 0.416 0.458 0.465 0.483 ...

Gini.b.a.order.melt의 Country.order가 개선도 올림차순으로 정리되어 있는 factor이기 때문에 그대로 활용하면 됨.

ggplot(Gini.b.a.order.melt, aes(x = Country.order, y = Gini.Coef, fill = Tax)) + geom_bar(stat = "identity", position = "dodge") +
coord_flip()

한글 제목 등의 세부 작업은 차후에

뒷 마무리

save(file="Gini_OECD1012.rda", list=ls())
# savehistory("Gini_OECD102.Rhistory")

Gini_OECD

coop711

2015년 5월 4일

Data 작업

Graphic representation

ggplot

Data reshaping

뒷 마무리