R & Twitter: statistics and word-cloud

Twitter Analysis with R

Анализ коротких сообщений в Twitter на основе R

Connect to Twitter

consumerKey"55ZCRbatWWVYpr2VU9Ddk1234"

consumerSecret"ji4Zqit6jL7iIoegHKr3zVBoCTEeYSnmH0aAJdm3j3gRZRFlL7"
setup_twitter_oauth(consumer_key=consumerKey,consumer_secret=consumerSecret)

Search Tweets

tweets1 = searchTwitter("business analytics", n=250)

Twitter List to Data Frame

data1 = twListToDF(tweets1)

Plots

qplot(created, data = data1, geom = "bar",weight= retweetCount)

and/or

cc1 = ggplot(data1, aes(created))
cc1 + geom_bar()

or

cc1 = ggplot(data1, aes(created))
cc1 + geom_histogram()

and/or

ggplot(data1, aes(data1$created,data1$retweetCount)) + geom_line()

Label

data1$month = sapply(data1$created, function(x) {p=as.POSIXlt(x);p$mon})
data1$hour = sapply(data1$created, function(x) {p=as.POSIXlt(x);p$hour})
data1$wday = sapply(data1$created, function(x) {p=as.POSIXlt(x);p$wday})
data1$year = sapply(data1$created, function(x) {p=as.POSIXlt(x);p$year})

Make More Diagrams

ggplot(data1,aes(x=wday))+geom_histogram(aes(y = (..count..)),binwidth = 1)

and/or

ggplot(data4)+geom_bar(aes(x=wday))

and/or

ggplot(data1)+geom_jitter(aes(x=wday,y=hour))

Word Cloud

Prepare Text

text1 = sapply(tweets1, function(x) x$getText())

text1 = tolower(text1)

text1 = gsub("rt", "", text1);

text1 = gsub("@\\w+", "", text1)

text1 = gsub("[[:punct:]]", "", text1)

text1 = gsub("http\\w+", "", text1)

text1 = gsub("[ |\t]{2,}", "", text1)

text1 = gsub("^ ", "", text1)

text1 = gsub(" $", "", text1)

Make Word Cloud

wordcloud(text1, min.freq = 1, scale=c(7,0.5), colors=brewer.pal(8, "Dark2"),  random.color= TRUE, random.order = FALSE, max.words = 150)

Полный текст статьи, посвященный вопросам анализа коротких сообщений в Twitter и создании на их основе облака слов, находится здесь.