-
Notifications
You must be signed in to change notification settings - Fork 0
/
twitter_analysis.R
302 lines (174 loc) · 9.49 KB
/
twitter_analysis.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
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
library(twitteR)
library(lubridate)
library(ggplot2)
library(tm)
library(wordcloud)
consumerKey='x'
consumerSecret='x'
accesstoken ='x'
tokensecret = 'x'
#establishing connection with Twitter api
setup_twitter_oauth(consumerKey,
consumerSecret,
accesstoken,
tokensecret)
#searchTwitter gives only a week's data
#cdTweets <- searchTwitter('from:cardekho', n=5000 ,since = '2014-09-01', until = '2015-10-27')
cdTweets2 <- userTimeline('cardekho',n=3200,includeRts = T)
cdtwt2 <- twListToDF(cdTweets2)
View(cdtwt2)
setwd('E:\\DM\\cardekho')
#write.csv(cdtwt2,'cardekho_tweets.csv',row.names = F)
#cdtwt2 <- read.csv('cardekho_tweets.csv',header = T,stringsAsFactors = F)
#extracting one-year data
cardekho <- cdtwt2[1:2220,]
#saving a copy
write.csv(cardekho,'cardekho_twt1yr.csv',row.names = F)
View(cardekho)
cardekho <- read.csv('cardekho_twt1yr.csv',header = T,stringsAsFactors = F)
#removing extra columns
cardekho <- cardekho[,c(1,3,4,5,12,13)]
#date-time changing timezone to IST
cardekho$dt <- with_tz(ymd_hms(cardekho$created),'Asia/Calcutta')
#calculating time difference between two tweets
for(i in 1:2219){
cardekho$timedif[i] = round(as.numeric(difftime(cardekho$dt[i],cardekho$dt[i+1],units='mins')),2)
cardekho$timedif[2220] = 0
}
#cardekho <- read.csv('cardekho_update2.csv',header = T)
cardekho$weekday = weekdays(cardekho$dt)
cardekho$weekday = as.factor(cardekho$weekday)
print(levels(cardekho$weekday))
cardekho$weekday = factor(cardekho$weekday,levels(cardekho$weekday)[c(4,2,6,7,5,1,3)])
#extracting time
cardekho$time = hms(sapply(strsplit(as.character(cardekho$dt)," "),'[',2))
#labeling time period of day
cardekho$tinterval[cardekho$time>hms('05:00:01')&cardekho$time<hms('09:30:00')] = 'morning'
cardekho$tinterval[cardekho$time>hms('09:30:01')&cardekho$time<hms('11:30:00')]= 'forenoon'
cardekho$tinterval[cardekho$time>hms('11:30:01')&cardekho$time<hms('13:30:00')] = 'midday'
cardekho$tinterval[cardekho$time>hms('13:30:01')&cardekho$time<hms('16:00:00')]= 'afternoon'
cardekho$tinterval[cardekho$time>hms('16:00:01')&cardekho$time<hms('19:00:00')] = 'evening'
cardekho$tinterval[cardekho$time>hms('19:00:01')&cardekho$time<hms('21:00:00')]= 'late evening'
cardekho$tinterval[cardekho$time>hms('21:00:01')] = 'night'
cardekho$tinterval[cardekho$time<hms('05:00:00')]= 'late night'
cardekho$tinterval = as.factor(cardekho$tinterval)
print(levels(cardekho$tinterval))
cardekho$tinterval = factor(cardekho$tinterval,levels(cardekho$tinterval)[c(7,3,6,1,2,4,8,5)])
#extracting original tweets from overall tweets
cardekho_og <- cardekho[cardekho$isRetweet=='FALSE',]
View(cardekho_og)
#visualization
x = as.data.frame(round(prop.table(table(cardekho$weekday)),2))
ggplot(x)+geom_bar(aes(x= x$Var1,y=x$Freq),stat = "identity") +
xlab('Weekdays') + ylab('Frequency') +
ggtitle('Amount of Tweets in a Week')
weekday_rtcount = aggregate(cardekho_og$retweetCount~cardekho_og$weekday,data=cardekho_og,FUN=mean)
ggplot(weekday_rtcount)+geom_bar(aes(x= weekday_rtcount$`cardekho_og$weekday`,y=weekday_rtcount$`cardekho_og$retweetCount`),stat = "identity") +
xlab('Day of the Week') + ylab('Avg. no. of RTs per tweet') +
ggtitle('RTs per tweet in a Week')
round(prop.table(table(cardekho$tinterval)),3)
y = as.data.frame(round(prop.table(table(cardekho$tinterval)),3))
ggplot(y)+geom_bar(aes(x= y$Var1,y=y$Freq),stat = "identity") +
xlab('Time in Day') + ylab('Frequency') +
ggtitle('Amount of Tweets in a Day')
tint_rtcount = aggregate(cardekho_og$retweetCount~cardekho_og$tinterval,data=cardekho_og,FUN=mean)
ggplot(tint_rtcount)+geom_bar(aes(x= tint_rtcount$`cardekho_og$tinterval`,y=tint_rtcount$`cardekho_og$retweetCount`),stat = "identity") +
xlab('Time Interval in the day') + ylab('Avg. no. of RTs per tweet') +
ggtitle('RTs per tweet')
ggplot(as.data.frame(cardekho$timedif[cardekho$timedif<1200])) + geom_histogram(aes(x=cardekho$timedif[cardekho$timedif<1200],fill = ..count..),binwidth = 5) +
xlab('Time difference between tweets(in mins)') +
ylab('No. of tweets') +
ggtitle('Frequency of Tweets')
ggplot(as.data.frame(cardekho$timedif[cardekho$timedif<120])) + geom_histogram(aes(x=cardekho$timedif[cardekho$timedif<120],fill = ..count..),binwidth = 5) +
xlab('Time difference between tweets(in mins)') +
ylab('No. of tweets') +
ggtitle('Frequency of Tweets (witin 2 hrs)')
#visualizing tweet vs RT
ggplot(cardekho) + geom_bar(aes(x=factor(1),y = ((..count..)/sum(..count..))*100,fill=factor(cardekho$isRetweet)),width=1) +
coord_polar(theta = 'y') +
xlab('Percentage') +
ylab(' ') +
guides(fill=guide_legend(title='Is Retweet?')) +
ggtitle('Original Tweets vs Retweets')
#extracting top-performed original tweets
cardekho_toppers <- cardekho_og[cardekho_og$retweetCount>4,]
#mining the content of top-performing tweets
txt <- Corpus(VectorSource(cardekho_toppers$text))
txt <- tm_map(txt,tolower)
txt <- tm_map(txt,removePunctuation)
txt <- tm_map(txt,removeNumbers)
stpwords <- c(stopwords('english'),'the','\n','us','cardekho')
txt <- tm_map(txt,removeWords,stpwords)
txt <- sapply(1:22, function(x){gsub("http\\w+ *", "", txt[x]$content)})
txt <- gsub("\n\\w+ *", "", txt)
wordcloud(txt,scale = c(3, 0.1),min.freq = 2,colors = c('blue','green','red'),random.color =T,max.words = 50)
cwTweets2 <- userTimeline('carwale',n=3200,includeRts = T)
cwtwt2 <- twListToDF(cwTweets2)
View(cwtwt2)
setwd('E:\\DM\\cardekho')
write.csv(cwtwt2,'carwale_tweets.csv',row.names = F)
cwtwt2 <- read.csv('carwale_tweets.csv',header = T,stringsAsFactors = F)
#extracting one-year data
carwale <- cwtwt2[1:2560,]
#saving a copy
write.csv(carwale,'carwale_twt1yr.csv',row.names = F)
View(carwale)
carwale <- read.csv('carwale_twt1yr.csv',header = T,stringsAsFactors = F)
#removing extra columns
carwale <- carwale[,c(1,3,4,5,12,13)]
#date-time changing timezone to IST
carwale$dt <- with_tz(ymd_hms(carwale$created),'Asia/Calcutta')
#calculating time difference between two tweets
for(i in 1:2559){
carwale$timedif[i] = round(as.numeric(difftime(carwale$dt[i],carwale$dt[i+1],units='mins')),2)
carwale$timedif[2560] = 0
}
carwale$weekday = weekdays(carwale$dt)
carwale$weekday = as.factor(carwale$weekday)
print(levels(carwale$weekday))
carwale$weekday = factor(carwale$weekday,levels(carwale$weekday)[c(4,2,6,7,5,1,3)])
x = as.data.frame(round(prop.table(table(carwale$weekday)),2))
ggplot(x)+geom_bar(aes(x= x$Var1,y=x$Freq),stat = "identity") +
xlab('Weekdays') + ylab('Frequency') +
ggtitle('Amount of Tweets in a Week - Carwale')
weekday_rtcount = aggregate(carwale$retweetCount~carwale$weekday,data=carwale,FUN=mean)
ggplot(weekday_rtcount)+geom_bar(aes(x= weekday_rtcount$`carwale$weekday`,y=weekday_rtcount$`carwale$retweetCount`),stat = "identity") +
xlab('Day of the Week') + ylab('Avg. no. of RTs per tweet') +
ggtitle('RTs per tweet in a Week - Carwale')
carwale$time = hms(sapply(strsplit(as.character(carwale$dt)," "),'[',2))
carwale$tinterval[carwale$time>hms('05:00:01')&carwale$time<hms('09:30:00')] = 'morning'
carwale$tinterval[carwale$time>hms('09:30:01')&carwale$time<hms('11:30:00')]= 'forenoon'
carwale$tinterval[carwale$time>hms('11:30:01')&carwale$time<hms('13:30:00')] = 'midday'
carwale$tinterval[carwale$time>hms('13:30:01')&carwale$time<hms('16:00:00')]= 'afternoon'
carwale$tinterval[carwale$time>hms('16:00:01')&carwale$time<hms('19:00:00')] = 'evening'
carwale$tinterval[carwale$time>hms('19:00:01')&carwale$time<hms('21:00:00')]= 'late evening'
carwale$tinterval[carwale$time>hms('21:00:01')] = 'night'
carwale$tinterval[carwale$time<hms('05:00:00')]= 'late night'
carwale$tinterval = as.factor(carwale$tinterval)
print(levels(carwale$tinterval))
carwale$tinterval = factor(carwale$tinterval,levels(carwale$tinterval)[c(7,3,6,1,2,4,8,5)])
write.csv(carwale,'carwale_update2.csv',row.names =F)
round(prop.table(table(carwale$tinterval)),3)
y = as.data.frame(round(prop.table(table(carwale$tinterval)),3))
ggplot(y)+geom_bar(aes(x= y$Var1,y=y$Freq),stat = "identity") +
xlab('Time in Day') + ylab('Frequency') +
ggtitle('Amount of Tweets in a Day - Carwale')
tint_rtcount = aggregate(carwale$retweetCount~carwale$tinterval,data=carwale,FUN=mean)
ggplot(tint_rtcount)+geom_bar(aes(x= tint_rtcount$`carwale$tinterval`,y=tint_rtcount$`carwale$retweetCount`),stat = "identity") +
xlab('Time Interval in the day') + ylab('Avg. no. of RTs per tweet') +
ggtitle('RTs per tweet - Carwale')
ggplot(as.data.frame(carwale$timedif[carwale$timedif<1200])) + geom_histogram(aes(x=carwale$timedif[carwale$timedif<1200],fill = ..count..),binwidth = 5) +
xlab('Time difference between tweets(in mins)') +
ylab('No. of tweets') +
ggtitle('Frequency of Tweets - Carwale')
ggplot(as.data.frame(carwale$timedif[carwale$timedif<120])) + geom_histogram(aes(x=carwale$timedif[carwale$timedif<120],fill = ..count..),binwidth = 5) +
xlab('Time difference between tweets(in mins)') +
ylab('No. of tweets') +
ggtitle('Frequency of Tweets (witin 2 hrs) - Carwale')
#visualizing no. of retweets
ggplot(carwale) + geom_bar(aes(x=factor(1),y = ((..count..)/sum(..count..))*100,fill=factor(carwale$isRetweet)),width=1) +
coord_polar(theta = 'y') +
xlab('Percentage') +
ylab(' ') +
guides(fill=guide_legend(title='Is Retweet?')) +
ggtitle('Original Tweets vs Retweets - Carwale')