-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathMovie_dataset_cleaning
199 lines (165 loc) · 8.56 KB
/
Movie_dataset_cleaning
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
192
193
194
195
196
197
198
199
library(tidyr)
library(dplyr)
library(stringr)
library(compare)
library(ggplot2)
# Get the data file
raw_movie_data=read.csv("https://raw.githubusercontent.com/tressythomas/R_Learn/master/movie_dataset.csv",header=TRUE,sep=',')
#check the type of data, dimensions, structure and summary
class(raw_movie_data)
dim(raw_movie_data)
names(raw_movie_data)
glimpse(raw_movie_data)
summary(raw_movie_data)
#Rearrange the columns, drop unwanted columns - actor_1_facebook_likes ...facenumber_in_poster
field_order=c("movie_title","title_year","duration","language","country","color","budget","gross","genres","content_rating","director_name","actor_1_name","actor_2_name","actor_3_name","aspect_ratio","plot_keywords","num_critic_for_reviews","num_user_for_reviews","num_voted_users","imdb_score","movie_facebook_likes","cast_total_facebook_likes","director_facebook_likes","movie_imdb_link")
movie_data_1=raw_movie_data[,field_order]
#view first 6 rows
head(movie_data_1)
#Remove duplicates
movie_data_2=distinct(movie_data_1)
dim(movie_data_2)
#Remove rows with more than 10 columns with NA
#movie_data_2w=movie_data_2[which(rowMeans(!is.na(movie_data_2)) > 0.3), ]
#dim(movie_data_2w)
#check the NA fields in data frame
summary(is.na(movie_data_2))
#check if there is any blanks or NA for movie title
movie_data_2[is.na(movie_data_2$movie_title),]
#Clean the movie_title field. Trim trailing whitespace and Remove the last character. No NA.
movie_data_2$movie_title=as.character(movie_data_2$movie_title)
movie_data_2$movie_title=str_trim(movie_data_2$movie_title,"both")
movie_data_2$movie_title=substr(movie_data_2$movie_title,1,nchar(movie_data_2$movie_title)-1)
mva=movie_data_2[!duplicated(movie_data_2[c("movie_title","title_year","language"), ])]
#change the special characters
movie_data_2$movie_title=factor(iconv(movie_data_2$movie_title,from = 'UTF-8', to = 'ASCII//TRANSLIT'))
#-------Clean up Movie Year-------#
#histogram
hist(movie_data_2$title_year,
main='Movie Year Histogram',
xlab='Movie Year',
xlim=c(1920,2020),
ylim=c(0,1300),
col='lightblue',
breaks=20)
#Check the movies with year NOT present.
subset(movie_data_2,is.na(movie_data_2$title_year))
#Replace the NA year fields with default 1888 = Year Not Available.
movie_data_2$title_year[is.na(movie_data_2$title_year)]=1888
movie_data_3=data.frame(movie_data_2)
#-------Clean up Movie duration ----------#
#boxplot of duration
boxplot(movie_data_3$duration,
main='Movie Duration Boxplot',
xlab='Movie duration')
#view outlier and NA rows
summary(movie_data_2$duration)
movie_data_3$duration=as.integer(movie_data_3$duration)
dim(subset(movie_data_3,movie_data_3$duration < 10 | movie_data_3$duration > 300 | (is.na(movie_data_3$duration))))
#Replace the rows with NA duration to median and round the value for 10 and 300 for < 10 or > 300
movie_data_3$duration = ifelse(movie_data_3$duration < 10 , 10, movie_data_3$duration)
movie_data_3$duration = ifelse(movie_data_3$duration > 300 , 300, movie_data_3$duration)
movie_data_3$duration = ifelse(is.na(movie_data_3$duration), median(movie_data_3$duration,na.rm=TRUE), movie_data_3$duration)
summary(movie_data_3$duration)
movie_data_4=data.frame(movie_data_3)
#------- Clean up Language ---------------#
ggplot(data.frame(movie_data_4), aes(x=language)) +
geom_bar() +
theme(axis.text.x=element_text(angle=90,hjust=1))
summary(movie_data_4$language)
movie_data_4$language=str_trim(movie_data_2$language,"both")
subset(movie_data_4,movie_data_4$language=="")
#movie_data_2$language = ifelse(movie_data_2$language='' | movie_data_2$country = 'USA' , 'English', movie_data_2$language)
#Replace blank language fields with English
movie_data_4$language = sub("^$","English",movie_data_4$language)
movie_data_4$language = factor(movie_data_4$language)
summary(movie_data_4$language)
movie_data_5=data.frame(movie_data_4)
##-------Country----------#
summary(movie_data_5$country)
ggplot(data.frame(movie_data_5), aes(x=country)) +
geom_bar() +
theme(axis.text.x=element_text(angle=90,hjust=1))
#Drop the rows with country not present... No relevant info present in other columns also.
movie_data_5=movie_data_5[!(movie_data_3$country == ""),]
summary(movie_data_5$country)
movie_data_6=data.frame(movie_data_5)
##------- Clean Up Color----------#
summary(movie_data_6$color)
movie_data_6$color=str_trim(movie_data_6$color,"both")
movie_data_6[(movie_data_6$color == ""),]
#Replace blanks with color
movie_data_6$color = sub("^$","Color",movie_data_6$color)
movie_data_6$color = factor(movie_data_6$color)
summary(movie_data_6$color)
movie_data_7=data.frame(movie_data_6)
###------- Clean up Budget and gross----------#
dim(movie_data_7[(is.na(movie_data_7$budget)) & (is.na(movie_data_7$gross)),])
plot(movie_data_4$budget,movie_data_4$gross,
xlab="budget", ylab = "gross", pch=90,
main="Budget Gross Plot")
boxplot(movie_data_4$gross,
xlab="gross",
main="gross Box Plot")
#Replace NA with 0
movie_data_7$budget=ifelse(is.na(movie_data_7$budget),0,movie_data_7$budget)
movie_data_7$gross=ifelse(is.na(movie_data_7$gross),0,movie_data_7$gross)
#Remove outlier
movie_data_7=subset(movie_data_7,movie_data_7$budget < 12215500000)
movie_data_8=data.frame(movie_data_7)
##------- Clean up Genre ------- # No cleaning
summary(movie_data_8$genres)
##------- Clean up Content Rating
summary(movie_data_8$content_rating)
class(movie_data_8$content_rating)
movie_data_8$content_rating=sub("^$","Not Rated",movie_data_8$content_rating)
movie_data_8$content_rating=factor(movie_data_8$content_rating)
movie_data_9=data.frame(movie_data_8)
summary(movie_data_9)
#--------clean up Director and actor names, replace blanks with Unknown
movie_data_9$director_name=sub("^$","Unknown",movie_data_9$director_name)
movie_data_9$director_name=factor(iconv(movie_data_9$director_name, from = 'UTF-8', to = 'ASCII//TRANSLIT'))
#concatenate actors to a single column
movie_data_9$actors=paste(movie_data_9$actor_1_name,movie_data_9$actor_2_name,movie_data_9$actor_3_name,sep=' ')
movie_data_9$actors=str_trim(movie_data_9$actors,"right")
movie_data_9$actors=sub(" ", "|", movie_data_9$actors)
movie_data_9$actors=sub(" ", "|", movie_data_9$actors)
movie_data_9$actors=sub("^$","Unknown",movie_data_9$actors)
movie_data_9$actors=factor(iconv(movie_data_9$actors, from = 'UTF-8', to = 'ASCII//TRANSLIT'))
head(movie_data_9)
movie_data_10=data.frame(movie_data_9)
#aspect ratio
hist(movie_data_10$aspect_ratio,main="Aspect Ratio Histogram")
boxplot(movie_data_10$aspect_ratio,main="Aspect Ratio boxplot")
#replace 16 with 1.6 and NA with median
movie_data_10$aspect_ratio=ifelse(movie_data_10$aspect_ratio==16,1.6,movie_data_10$aspect_ratio)
movie_data_10$aspect_ratio=ifelse(is.na(movie_data_10$aspect_ratio),1.85,movie_data_10$aspect_ratio)
#plot keywords- can be empty
#num_critic_for_reviews, num_user_for_reviews
hist(movie_data_4$num_critic_for_reviews)
#replace na with 0
movie_data_10$num_critic_for_reviews=ifelse(is.na(movie_data_10$num_critic_for_reviews),0,movie_data_10$num_critic_for_reviews)
movie_data_10$num_user_for_reviews=ifelse(is.na(movie_data_10$num_user_for_reviews),0,movie_data_10$num_user_for_reviews)
#format imbd score
movie_data_10$imdb_score=as.numeric(movie_data_10$imdb_score)
#movie_data_10$imdb_score=format(round(movie_data_10$imdb_score, 1), nsmall = 1)
#movie_facebook_likes, cast_total_facebook_likes - no change
#director_facebook_likes
dim(movie_data_10[(is.na(movie_data_10$director_facebook_likes)),])
movie_data_10$director_facebook_likes=ifelse(is.na(movie_data_10$director_facebook_likes),0,movie_data_10$director_facebook_likes)
#Rename Column names
summary(is.na(movie_data_10))
names(movie_data_10)
field_order=c("movie_title","title_year","duration","language","country","color","budget","gross","genres","content_rating","director_name","actors","aspect_ratio","plot_keywords","num_critic_for_reviews","num_user_for_reviews","imdb_score","movie_facebook_likes","cast_total_facebook_likes","director_facebook_likes","movie_imdb_link")
movie_data_work=movie_data_10[,field_order]
dim(movie_data_10)
dim(movie_data_work)
names(movie_data_work)
col_names=c("movie_title","movie_release_year","movie_runtime","language","country","color","budget","gross","genres","content_rating","director_name","actors","aspect_ratio","plot_keywords","critic_reviews","user_reviews","imdb_score","movie_facebook_likes","cast_total_facebook_likes","director_facebook_likes","movie_imdb_link")
colnames(movie_data_work)=col_names
head(movie_data_work)
summary(movie_data_work)
movie_data_final=data.frame(movie_data_work)
summary(is.na(movie_data_final))
dim(movie_data_final)
write.csv(movie_data_final,'movie_dataset_final.csv')