#10 .기술통계 [ 예제 ]

예제로 개념 잡기 

DF <- read.csv("data/example_salary.csv",stringsAsFactors = T,na="-") # na is "-"

head(DF)

#    age  salary specialSalary workingTime numberOfWorker  career sex       scale

# 1   -19 1346534        151840       169.5          15042 1년미만  남 -1.28886999

# 2 20-24 1584214        115375       180.1          74251 1년미만  남 -0.91757018

# 3 25-29 1922043        268058       178.0         143338 1년미만  남 -0.38981924

# 4 30-34 2130988        335710       180.8         103585 1년미만  남 -0.06340878

# 5 35-39 2414345        352816       181.4          65385 1년미만  남  0.37924689

# 6 40-44 2372214        233728       182.9          55422 1년미만  남  0.31343053

summary(DF)

colnames(DF) <- c("age","salary","specialSalary","workingTime","numberOfWorker","career","sex")

detach(DF)

attach(DF)

Mean <- mean(salary,na.rm=T)    

Mid <- median(salary,na.rm=T)

Range <- range(salary,na.rm=T)

w <- which(DF$salary==4064286)

DF[w,]

Qnt <- quantile(salary,na.rm=T)

SelfSummary <- function(x){

  Mean <- mean(salary,na.rm=T)

  Mid <- median(salary,na.rm=T)

  Range <- range(salary,na.rm=T)

  Qnt <- quantile(salary,na.rm=T)

  Result <- list(평균월급=Mean,중앙값월급=Mid,월급번위=Range, 월급사분위 = Qnt)

  return(Result)

}

Result <- SelfSummary(salary)

Result

$평균월급

[1] 2171578

$중앙값월급

[1] 2120345

$월급번위

[1] 1117605 4064286

$월급사분위

     0%     25%     50%     75%    100% 

1117605 1689658 2120345 2519221 4064286 

############## 그룹별 평균구하기   #############

temp <- tapply(salary,sex, mean,na.rm=T)   # tapply(데이터,기준,함수)

#     남      여 

# 2477332 1865823 

class(temp) # array

library(reshape2)

library(ggplot2)

melt <- melt(temp)  # array = > dataFrame 

#  Var1   value

# 1   남 2477332

# 2   여 1865823

ggplot(melt,aes(x=Var1,y=value,fill=Var1)) + geom_bar(stat="identity")

# 성별 별 표준편차

tapply(salary, sex, sd, na.rm=T)

#      남       여 

# 646470.7 468270.6 

# 성별 별 월급 범위.

tapply(salary, sex, range, na.rm=T)

# 경력별로 월급의 평균

tapply(salary,career, mean, na.rm=T)

melt <- melt(tapply(salary,career, mean, na.rm=T))

ggplot(melt,aes(x=Var1,y=value,group=1)) + geom_line(colour="skyblue2",size=2) + coord_polar() + ylim(0,max(melt$value))

tapply(salary, career,sd,na.rm=T)

temp <- tapply(salary, career,range,na.rm=T)

temp[[1]][1] # 1~3

temp[[2]][1] # 10년 이상

temp[[3]][1] # 1년미만

temp[[4]][1] # 3~5년미만

temp[[5]][1] # 5~10년미만

a1 <- DF[which(salary==temp[[1]][1]),]

a2 <- DF[which(salary==temp[[2]][1]),]

a3 <- DF[which(salary==temp[[3]][1]),]

a4 <- DF[which(salary==temp[[4]][1]),]

a5 <- DF[which(salary==temp[[5]][1]),]

list <- list(a1,a2,a3,a4,a5)

# [[1]]

# age  salary specialSalary workingTime numberOfWorker    career sex

# 70 60- 1172399        299639       151.2          30253 1~3년미만  여

# 

# [[2]]

# age  salary specialSalary workingTime numberOfWorker   career sex

# 92 20-24 1685204       1970720       179.4           1886 10년이상  여

# 

# [[3]]

# age  salary specialSalary workingTime numberOfWorker  career sex

# 60 60- 1117605         10667         148          18737 1년미만  여

# 

# [[4]]

# age  salary specialSalary workingTime numberOfWorker    career sex

# 80 60- 1245540        423826       155.1          21106 3~5년미만  여

# 

# [[5]]

# age  salary specialSalary workingTime numberOfWorker     career sex

# 90 60- 1548036        806919       169.1          20282 5~10년미만  여

############# IQR 구하기 #################

Salary <- summary(salary,na.rm=T)

distIQR <- Salary[[5]] - Salary[[2]]

iqr <- IQR(salary,na.rm=T)

############ 아웃라이어 찾기와 제거 ######

DF <- read.csv("data/example_cancer.csv",stringsAsFactors = F,na="기록없음")

str(DF)

# 'data.frame': 18310 obs. of  8 variables:

#   $ age            : int  75 52 67 62 70 76 55 72 64 71 ...

# $ sex            : chr  "남" "여" "여" "남" ...

# $ height         : num  161 177 154 162 171 ...

# $ weight         : num  64 75.3 65.6 57 65 87 77 55 67 55.5 ...

# $ dateOfoperation: chr  "2011-06-22" "2011-05-19" "2011-05-31" "2011-06-21" ...

# $ cancerStaging  : chr  "I" "IV" "III" "I" ...

# $ hospitalization: int  48 17 10 11 10 10 12 18 15 35 ...

# $ diseaseCode    : chr  "C187" "C187" "C187" "C187" ...

detach(DF)

attach(DF)

mean(age)

summary(age)

boxplot(age)  # 대략적으로 30대 이하의 사람들이 Outlier 또는 100세 이상. 

grid() #모눈 그리기

distIQR <- IQR(age, na.rm=T)

# 17

posIQR <- quantile(age,probs=c(0.25,0.75),na.rm=T)

# 25% 75% 

#  55  72 

DownWhisker <- posIQR[[1]] - distIQR * 1.5 # 25%지점에서 1.5배만큼 연장.

UpWhisker <- posIQR[[2]] + distIQR * 1.5

Outlier <- subset(DF, subset = (DF$age < DownWhisker | DF$age > UpWhisker))

########### 평균값 표준화하여 그래프를 그리기. ( 표준화 그래프 )

library(ggplot2)

DF <- read.csv("data/example_salary.csv",stringsAsFactors = F, na="-")

colnames(DF) <- c("age","salary","specialSalary","workingTime","numberOfWorker","career","sex")

Scale <- scale(DF$salary)

# Scale <- matrix(Scale[!is.na(Scale)])

DF <- cbind(DF,scale=Scale)

DF2 <- DF[!is.na(DF$scale),]

g1 <- ggplot(DF,aes(x=scale,y=age))

g2 <- geom_segment(aes(yend=age),xend=0)

g3 <- g1 + g2 + geom_point(size=7, aes(colour=sex,shape=career)) + theme_minimal() + theme(text = element_text(size=20))

g3