/
yelp.R
208 lines (168 loc) · 7.88 KB
/
yelp.R
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
200
201
202
203
###########################################
# R commands to process the Yelp database #
###########################################
#############################################
# Part 1: Setup and initial data wrangling #
#############################################
# Load library
library(dplyr)
# Read in csv files
reviews <- read.csv("yelp_academic_dataset_review.csv", header = FALSE)
users <- read.csv("yelp_academic_dataset_user.csv", header = FALSE)
businesses <- read.csv("yelp_academic_dataset_business.csv", header = FALSE)
# Add names to the fields
colnames(reviews)[1] = "user_id"
colnames(reviews)[2] = "business_id"
colnames(reviews)[3] = "stars"
colnames(users)[1] = "user_id"
colnames(users)[2] = "user_name"
colnames(businesses)[1] = "business_id"
colnames(businesses)[2] = "city"
colnames(businesses)[3] = "business_name"
colnames(businesses)[4] = "categories"
colnames(businesses)[5] = "review_count"
colnames(businesses)[6] = "avg_stars"
# Join the files
ru <- inner_join(reviews, users)
rub <- inner_join(ru, businesses)
######################################################
# Part 2a: Analysis of Method 1 -- Initial Analysis #
######################################################
# Add "is_indian" field for any review that has "Indian" in "categories"
rub$is_indian <- grepl("Indian", rub$categories) == TRUE
# Make a dataframe of just reviews of Indian restaurants
indian <- subset(rub, is_indian == TRUE)
# Generate a summary of # of reviews of that cuisine done by each reviewer
num_reviews_Indian <- indian %>% select(user_id, user_name, is_indian) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_indian))
# Print the table, show the total # of entries, and find the avg # of reviews per user
table(num_reviews_Indian$tot_rev)
count(num_reviews_Indian)
mean(num_reviews_Indian$tot_rev)
#################################################################
# Part 2b: Analysis of Method 1 -- Extension to Other Cuisines #
#################################################################
rub$is_chinese <- grepl("Chinese", rub$categories) == TRUE
chinese <- subset(rub, is_chinese == TRUE)
num_reviews_Chinese <- chinese %>% select(user_id, user_name, is_chinese) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_chinese))
table(num_reviews_Chinese$tot_rev)
count(num_reviews_Chinese)
mean(num_reviews_Chinese$tot_rev)
rub$is_mexican <- grepl("Mexican", rub$categories) == TRUE
mexican <- subset(rub, is_mexican == TRUE)
num_reviews_Mexican <- mexican %>% select(user_id, user_name, is_mexican) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_mexican))
table(num_reviews_Mexican$tot_rev)
count(num_reviews_Mexican)
mean(num_reviews_Mexican$tot_rev)
rub$is_italian <- grepl("Italian", rub$categories) == TRUE
italian <- subset(rub, is_italian == TRUE)
num_reviews_Italian <- italian %>% select(user_id, user_name, is_italian) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_italian))
table(num_reviews_Italian$tot_rev)
count(num_reviews_Italian)
mean(num_reviews_Italian$tot_rev)
# For Japanese, look for "Japanese" or "Sushi"
rub$is_japanese <- (grepl("Japanese", rub$categories) == TRUE) |
(grepl("Sushi", rub$categories) == TRUE)
japanese <- subset(rub, is_japanese == TRUE)
num_reviews_Japanese <- japanese %>% select(user_id, user_name, is_japanese) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_japanese))
table(num_reviews_Japanese$tot_rev)
count(num_reviews_Japanese)
mean(num_reviews_Japanese$tot_rev)
rub$is_greek <- grepl("Greek", rub$categories) == TRUE
greek <- subset(rub, is_greek == TRUE)
num_reviews_Greek <- greek %>% select(user_id, user_name, is_greek) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_greek))
table(num_reviews_Greek$tot_rev)
count(num_reviews_Greek)
mean(num_reviews_Greek$tot_rev)
rub$is_french <- grepl("French", rub$categories) == TRUE
french <- subset(rub, is_french == TRUE)
num_reviews_French <- french %>% select(user_id, user_name, is_french) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_french))
table(num_reviews_French$tot_rev)
count(num_reviews_French)
mean(num_reviews_French$tot_rev)
rub$is_thai <- grepl("Thai", rub$categories) == TRUE
thai <- subset(rub, is_thai == TRUE)
num_reviews_Thai <- thai %>% select(user_id, user_name, is_thai) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_thai))
table(num_reviews_Thai$tot_rev)
count(num_reviews_Thai)
mean(num_reviews_Thai$tot_rev)
rub$is_spanish <- (grepl("Spanish", rub$categories) == TRUE) |
(grepl("Tapas", rub$categories) == TRUE)
spanish <- subset(rub, is_spanish == TRUE)
num_reviews_Spanish <- spanish %>% select(user_id, user_name, is_spanish) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_spanish))
table(num_reviews_Spanish$tot_rev)
count(num_reviews_Spanish)
mean(num_reviews_Spanish$tot_rev)
rub$is_mediterranean <- grepl("Mediterranean", rub$categories) == TRUE
mediterranean <- subset(rub, is_mediterranean == TRUE)
num_reviews_Mediterranean <- mediterranean %>% select(user_id, user_name, is_mediterranean) %>%
group_by(user_id) %>%
summarise(tot_rev = sum(is_mediterranean))
table(num_reviews_Mediterranean$tot_rev)
count(num_reviews_Mediterranean)
mean(num_reviews_Mediterranean$tot_rev)
#####################################################################
# Part 2c: Analysis of Method 1 -- Apply new weight and see effect #
#####################################################################
# Combine num_reviews information with original data frame of indian restaurant reviews
cin <- inner_join(indian, num_reviews_Indian)
# Generate "weighted_stars" for later calculation
cin$weighted_stars <- cin$stars * cin$tot_rev
# Use "summarise" to generate a new rating for each restaurant
new_rating_Indian <- cin %>% select(city, business_name, avg_stars, stars,
tot_rev, weighted_stars) %>%
group_by(city, business_name, avg_stars) %>%
summarise(cnt = n(),
avg = sum(stars) / cnt,
new = sum(weighted_stars) / sum(tot_rev),
dif = new - avg)
# Print summary data of the effect this new rating has
summary(new_rating_Indian$dif)
# Limit to those with at least 5 ratings and redo summary
nri5 <- subset(new_rating_Indian, cnt > 5)
summary(nri5$dif)
################################################################
# Part 3: Analysis of Method 2 -- Generate "immigrant" rating #
################################################################
# Read Indian names into a list
inames <- scan("indian_names.txt", what = character())
# Add field "reviewer_indian_name" to indian reviews if user name is in the list
indian$reviewer_indian_name <- indian$user_name %in% inames
# Generate "istars" for internal calculation later
indian$istars <- indian$stars * indian$reviewer_indian_name
# Find out # of reviewers with a uniquely Indian name
table(indian$reviewer_indian_name)
1274/(1274 + 11872) # .096
# Generate new "immigrant" rating
avg_rating_Indian <- indian %>% select(business_id, business_name, city, stars,
avg_stars, reviewer_indian_name,
is_indian, istars) %>%
group_by(city, business_name, avg_stars) %>%
summarise(count = n(),
nin = sum(reviewer_indian_name),
pin = sum(reviewer_indian_name) / n(),
avg = sum(stars) / count,
ias = sum(istars) / nin,
dif = ias - avg)
# Find out extent of effect of new rating
summary(avg_rating_Indian$dif)
# Limit to those restaurants with at least 5 "immigrant" reviews and look at effect again
ari5 <- subset (avg_rating_Indian, nin > 5)
summary(ari5$dif)