#上課程式碼：
#資料介紹：https://www.kaggle.com/hb20007/gender-classification?select=Transformed+Data+Set+-+Sheet1.csv
#檔名為"Transformed Data Set - Sheet1.csv"
#從電腦端上傳(upload)到rstudio cloud
#install.packages("readr")#第一次使用需安裝
library(readr)#用library呼叫套件
#讀取資料，命名為"x"
x <- read.csv("Transformed Data Set - Sheet1.csv", stringsAsFactors = TRUE)
x
## Favorite.Color Favorite.Music.Genre Favorite.Beverage Favorite.Beverage.1
## 1 Cool Rock Vodka Vodka
## 2 Neutral Hip hop Vodka Vodka
## 3 Warm Rock Wine Wine
## 4 Warm Folk/Traditional Whiskey Whiskey
## 5 Cool Rock Vodka Vodka
## 6 Warm Jazz/Blues Doesn't drink Doesn't drink
## 7 Cool Pop Beer Beer
## 8 Warm Pop Whiskey Whiskey
## 9 Warm Rock Other Other
## 10 Neutral Pop Wine Wine
## 11 Cool Pop Other Other
## 12 Warm Pop Other Other
## 13 Warm Pop Wine Wine
## 14 Warm Electronic Wine Wine
## 15 Cool Rock Beer Beer
## 16 Warm Jazz/Blues Wine Wine
## 17 Cool Pop Wine Wine
## 18 Cool Rock Other Other
## 19 Cool Rock Other Other
## 20 Cool Pop Doesn't drink Doesn't drink
## 21 Cool Pop Beer Beer
## 22 Warm Jazz/Blues Whiskey Whiskey
## 23 Cool Rock Vodka Vodka
## 24 Warm Pop Other Other
## 25 Cool Folk/Traditional Whiskey Whiskey
## 26 Warm R&B and soul Whiskey Whiskey
## 27 Cool Pop Beer Beer
## 28 Cool Pop Doesn't drink Doesn't drink
## 29 Cool Pop Doesn't drink Doesn't drink
## 30 Cool Electronic Doesn't drink Doesn't drink
## 31 Warm Rock Other Other
## 32 Neutral Rock Beer Beer
## 33 Cool R&B and soul Beer Beer
## 34 Warm R&B and soul Wine Wine
## 35 Neutral Hip hop Beer Beer
## 36 Warm Electronic Other Other
## 37 Neutral Rock Doesn't drink Doesn't drink
## 38 Cool Pop Other Other
## 39 Cool Pop Whiskey Whiskey
## 40 Warm Rock Vodka Vodka
## 41 Cool Rock Vodka Vodka
## 42 Neutral Pop Doesn't drink Doesn't drink
## 43 Warm R&B and soul Doesn't drink Doesn't drink
## 44 Cool Rock Wine Wine
## 45 Cool Folk/Traditional Beer Beer
## 46 Cool Hip hop Beer Beer
## 47 Cool Hip hop Wine Wine
## 48 Cool R&B and soul Whiskey Whiskey
## 49 Cool Rock Doesn't drink Doesn't drink
## 50 Warm Hip hop Beer Beer
## 51 Cool R&B and soul Doesn't drink Doesn't drink
## 52 Cool Rock Doesn't drink Doesn't drink
## 53 Cool Hip hop Doesn't drink Doesn't drink
## 54 Warm Rock Beer Beer
## 55 Cool Electronic Doesn't drink Doesn't drink
## 56 Cool Electronic Other Other
## 57 Warm Folk/Traditional Other Other
## 58 Warm Electronic Vodka Vodka
## 59 Warm Jazz/Blues Vodka Vodka
## 60 Cool Pop Whiskey Whiskey
## 61 Cool Electronic Whiskey Whiskey
## 62 Cool Rock Vodka Vodka
## 63 Cool Hip hop Beer Beer
## 64 Neutral Hip hop Doesn't drink Doesn't drink
## 65 Cool Rock Wine Wine
## 66 Cool Electronic Beer Beer
## Gender
## 1 F
## 2 F
## 3 F
## 4 F
## 5 F
## 6 F
## 7 F
## 8 F
## 9 F
## 10 F
## 11 F
## 12 F
## 13 F
## 14 F
## 15 F
## 16 F
## 17 F
## 18 F
## 19 F
## 20 F
## 21 F
## 22 F
## 23 F
## 24 F
## 25 F
## 26 F
## 27 F
## 28 F
## 29 F
## 30 F
## 31 F
## 32 F
## 33 F
## 34 M
## 35 M
## 36 M
## 37 M
## 38 M
## 39 M
## 40 M
## 41 M
## 42 M
## 43 M
## 44 M
## 45 M
## 46 M
## 47 M
## 48 M
## 49 M
## 50 M
## 51 M
## 52 M
## 53 M
## 54 M
## 55 M
## 56 M
## 57 M
## 58 M
## 59 M
## 60 M
## 61 M
## 62 M
## 63 M
## 64 M
## 65 M
## 66 M
#查看資料結構
summary(x)
## Favorite.Color Favorite.Music.Genre Favorite.Beverage
## Cool :37 Electronic : 8 Beer :13
## Neutral: 7 Folk/Traditional: 4 Doesn't drink:14
## Warm :22 Hip hop : 8 Other :11
## Jazz/Blues : 4 Vodka : 9
## Pop :17 Whiskey : 9
## R&B and soul : 6 Wine :10
## Rock :19
## Favorite.Beverage.1 Gender
## Beer :13 F:33
## Doesn't drink:14 M:33
## Other :11
## Vodka : 9
## Whiskey : 9
## Wine :10
##
#重新命名欄位名稱
colnames(x) #檢視欄位名稱
## [1] "Favorite.Color" "Favorite.Music.Genre" "Favorite.Beverage"
## [4] "Favorite.Beverage.1" "Gender"
colnames(x) <- c("color", "music", "beverage", "drink", "gender")
###類別資料分析###
#問題：男女生各有多少人？
table(x$gender)#算次數
##
## F M
## 33 33
prop.table(table(x$gender))#算百分比
##
## F M
## 0.5 0.5
#問題：最喜歡的color、music、beverage、 drink、gender？
table(x$color)
##
## Cool Neutral Warm
## 37 7 22
prop.table(table(x$color))
##
## Cool Neutral Warm
## 0.5606061 0.1060606 0.3333333
#問題：男女對顏色的喜好的差異？
d <- table(x$gender,x$color)
d
##
## Cool Neutral Warm
## F 17 3 13
## M 20 4 9
#百分比次數分配表
p.d<- prop.table(d,2)
p.d
##
## Cool Neutral Warm
## F 0.4594595 0.4285714 0.5909091
## M 0.5405405 0.5714286 0.4090909
#將次數變成百分比(乘以100)
p.d <- p.d*100
p.d
##
## Cool Neutral Warm
## F 45.94595 42.85714 59.09091
## M 54.05405 57.14286 40.90909
#四捨五入至小數2位
p.d <- round(p.d,2)
p.d
##
## Cool Neutral Warm
## F 45.95 42.86 59.09
## M 54.05 57.14 40.91
#畫分組長條圖
barplot(p.d)
barplot(p.d, beside =T)
#加上圖例與上色
rownames(p.d)
## [1] "F" "M"
label <- rownames(p.d)
#次數分配表
t <- table(x$gender,x$drink)
t
##
## Beer Doesn't drink Other Vodka Whiskey Wine
## F 6 5 7 4 5 6
## M 7 9 4 5 4 4
#百分比次數分配表
p.t <- prop.table(t,1)
p.t
##
## Beer Doesn't drink Other Vodka Whiskey Wine
## F 0.1818182 0.1515152 0.2121212 0.1212121 0.1515152 0.1818182
## M 0.2121212 0.2727273 0.1212121 0.1515152 0.1212121 0.1212121
#將次數變成百分比(乘以100)
p.t <- p.t*100
p.t
##
## Beer Doesn't drink Other Vodka Whiskey Wine
## F 18.18182 15.15152 21.21212 12.12121 15.15152 18.18182
## M 21.21212 27.27273 12.12121 15.15152 12.12121 12.12121
#四捨五入至小數2位
p.t <- round(p.t,2)
p.t
##
## Beer Doesn't drink Other Vodka Whiskey Wine
## F 18.18 15.15 21.21 12.12 15.15 18.18
## M 21.21 27.27 12.12 15.15 12.12 12.12
#畫分組長條圖
barplot(p.t)
barplot(p.t, beside =T)
#加上圖例與上色
rownames(p.t)
## [1] "F" "M"
label <- rownames(p.t)
label
## [1] "F" "M"
barplot(p.t,
beside = T,
legend.text =label,
col = c (12,16))
#畫圓餅圖
p.d
##
## Cool Neutral Warm
## F 45.95 42.86 59.09
## M 54.05 57.14 40.91
f <- p.d[1,] # 女性資料
m <- p.d[2,] # 男性資料
f
## Cool Neutral Warm
## 45.95 42.86 59.09
m
## Cool Neutral Warm
## 54.05 57.14 40.91
# par()是圖形控制函數，mfrow = c(1,2)表示建立一個1x2的空間，用來呈現後續的圖
par( mfrow = c(1,2))
pie(f, main ="女生")
pie(m, main ="男生")
dev.off() #離開par()
## null device
## 1
#畫圓餅圖並加上資料標籤
pie_category <- colnames(p.d)
pie_category
## [1] "Cool" "Neutral" "Warm"
f_label <- paste(pie_category, f,"%", sep = "")
f_label
## [1] "Cool45.95%" "Neutral42.86%" "Warm59.09%"
m_label <- paste(pie_category, m,"%", sep = "")
m_label
## [1] "Cool54.05%" "Neutral57.14%" "Warm40.91%"
par(mfrow = c(1,2))# c(1,2)，表示建立一個1x2的空間，用來呈現後續的圖
pie(f, labels = f_label, main = "Female")
pie(m, labels = m_label, main = "Male" )
dev.off() #離開par()
## null device
## 1
p.t
##
## Beer Doesn't drink Other Vodka Whiskey Wine
## F 18.18 15.15 21.21 12.12 15.15 18.18
## M 21.21 27.27 12.12 15.15 12.12 12.12
f <- p.t[1,] # 女性資料
m <- p.t[2,] # 男性資料
f
## Beer Doesn't drink Other Vodka Whiskey
## 18.18 15.15 21.21 12.12 15.15
## Wine
## 18.18
m
## Beer Doesn't drink Other Vodka Whiskey
## 21.21 27.27 12.12 15.15 12.12
## Wine
## 12.12
# par()是圖形控制函數，mfrow = c(1,2)表示建立一個1x2的空間，用來呈現後續的圖
par( mfrow = c(1,2))
pie(f, main ="女生")
## Warning in title(main = main, ...): conversion failure on '女生' in
## 'mbcsToSbcs': dot substituted for <e5>
## Warning in title(main = main, ...): conversion failure on '女生' in
## 'mbcsToSbcs': dot substituted for <a5>
## Warning in title(main = main, ...): conversion failure on '女生' in
## 'mbcsToSbcs': dot substituted for <b3>
## Warning in title(main = main, ...): conversion failure on '女生' in
## 'mbcsToSbcs': dot substituted for <e7>
## Warning in title(main = main, ...): conversion failure on '女生' in
## 'mbcsToSbcs': dot substituted for <94>
## Warning in title(main = main, ...): conversion failure on '女生' in
## 'mbcsToSbcs': dot substituted for <9f>
pie(m, main ="男生")
## Warning in title(main = main, ...): conversion failure on '男生' in
## 'mbcsToSbcs': dot substituted for <e7>
## Warning in title(main = main, ...): conversion failure on '男生' in
## 'mbcsToSbcs': dot substituted for <94>
## Warning in title(main = main, ...): conversion failure on '男生' in
## 'mbcsToSbcs': dot substituted for <b7>
## Warning in title(main = main, ...): conversion failure on '男生' in
## 'mbcsToSbcs': dot substituted for <e7>
## Warning in title(main = main, ...): conversion failure on '男生' in
## 'mbcsToSbcs': dot substituted for <94>
## Warning in title(main = main, ...): conversion failure on '男生' in
## 'mbcsToSbcs': dot substituted for <9f>
dev.off() #離開par()
## null device
## 1
#畫圓餅圖並加上資料標籤
pie_category <- colnames(p.t)
pie_category
## [1] "Beer" "Doesn't drink" "Other" "Vodka"
## [5] "Whiskey" "Wine"
f_label <- paste(pie_category, f,"%", sep = "")
f_label
## [1] "Beer18.18%" "Doesn't drink15.15%" "Other21.21%"
## [4] "Vodka12.12%" "Whiskey15.15%" "Wine18.18%"
m_label <- paste(pie_category, m,"%", sep = "")
m_label
## [1] "Beer21.21%" "Doesn't drink27.27%" "Other12.12%"
## [4] "Vodka15.15%" "Whiskey12.12%" "Wine12.12%"
par(mfrow = c(1,2))# c(1,2)，表示建立一個1x2的空間，用來呈現後續的圖
pie(f, labels = f_label, main = "Female")
pie(m, labels = m_label, main = "Male" )
dev.off() #離開par()
## null device
## 1