Lecture9: Network representation & TwitteR
Network representation
文字単位のn-gram:英文
test1<-"I am innocent of the allegations made against me, Ghosn told the Tokyo District Court in English, adding that he has been wrongly accused and unfairly detained."
substr(test1,1,4)## [1] "I am"size=4
len<-nchar(test1)-size+1
ngramLst <- c()
for(i in 1:nchar(test1)){
ngramLst<-rbind(ngramLst,(substr(test1,i,i+size-1)))
}
head(ngramLst)## [,1]
## [1,] "I am"
## [2,] " am "
## [3,] "am i"
## [4,] "m in"
## [5,] " inn"
## [6,] "inno"単語単位のn-gram
準備：単語単位のリスト作成
wordLst <- strsplit(test1, "[[:space:]]|[[:punct:]]")
wordLst <- unlist(wordLst)
wordLst <- tolower(wordLst)
wordLst <- wordLst[wordLst != ""]
wordLst## [1] "i" "am" "innocent" "of" "the"
## [6] "allegations" "made" "against" "me" "ghosn"
## [11] "told" "the" "tokyo" "district" "court"
## [16] "in" "english" "adding" "that" "he"
## [21] "has" "been" "wrongly" "accused" "and"
## [26] "unfairly" "detained"単語単位のngram
size=3
start=1
wordLst[start:(start+size-1)]## [1] "i" "am" "innocent"strLst<-c()
len<-length(wordLst)-size+1
for(i in seq(1,len,size)) {
strLst<-rbind(strLst,wordLst[i:(i+size-1)])
}
strLst## [,1] [,2] [,3]
## [1,] "i" "am" "innocent"
## [2,] "of" "the" "allegations"
## [3,] "made" "against" "me"
## [4,] "ghosn" "told" "the"
## [5,] "tokyo" "district" "court"
## [6,] "in" "english" "adding"
## [7,] "that" "he" "has"
## [8,] "been" "wrongly" "accused"
## [9,] "and" "unfairly" "detained"ネットワーク描画
ペアデータの作成
一行のデータ
str<-strLst[1,]
pLst <- c()
for(i in 1:(length(str)-1)){
for(j in (i+1):length(str)){
tmp<-cbind(str[i],str[j])
pLst<-rbind(pLst,tmp)
}
}
pLst## [,1] [,2]
## [1,] "i" "am"
## [2,] "i" "innocent"
## [3,] "am" "innocent"netwkPairs.R（ペア単語取得用関数）のロード
source("netwkPairs.R")ペアデータの取得
getPairs(strLst[1,])## [,1] [,2]
## [1,] "i" "am"
## [2,] "i" "innocent"
## [3,] "am" "innocent"getPairs(strLst[2,])## [,1] [,2]
## [1,] "of" "the"
## [2,] "of" "allegations"
## [3,] "the" "allegations"getPairsLst(strLst)## [,1] [,2]
## [1,] "i" "am"
## [2,] "i" "innocent"
## [3,] "am" "innocent"
## [4,] "of" "the"
## [5,] "of" "allegations"
## [6,] "the" "allegations"
## [7,] "made" "against"
## [8,] "made" "me"
## [9,] "against" "me"
## [10,] "ghosn" "told"
## [11,] "ghosn" "the"
## [12,] "told" "the"
## [13,] "tokyo" "district"
## [14,] "tokyo" "court"
## [15,] "district" "court"
## [16,] "in" "english"
## [17,] "in" "adding"
## [18,] "english" "adding"
## [19,] "that" "he"
## [20,] "that" "has"
## [21,] "he" "has"
## [22,] "been" "wrongly"
## [23,] "been" "accused"
## [24,] "wrongly" "accused"
## [25,] "and" "unfairly"
## [26,] "and" "detained"
## [27,] "unfairly" "detained"ペア頻度表
pFreq<-getPairsFreq(strLst)
head(pFreq)## Term1 Term2 Freq
## 1 i am 1
## 2 i innocent 1
## 3 am innocent 1
## 4 of the 1
## 5 of allegations 1
## 6 the allegations 1igraphを利用した描画（無向グラフ）
library(igraph)##
## Attaching package: 'igraph'## The following objects are masked from 'package:stats':
##
## decompose, spectrum## The following object is masked from 'package:base':
##
## unionwng<-as.undirected(graph.data.frame(pFreq))
plot(wng)
テキストファイルからネットワーク図作成
filename <- "Ghosn.txt"
txt<-readLines(filename)
txt<- splitWdsEn(txt)
head(txt)## [1] "embattled" "former" "nissan" "motor" "co" "chairman"size=4
step=2
strLst<-getNstr(txt,size,step)
pFreq<-getPairsFreq(strLst)
head(pFreq)## Term1 Term2 Freq
## 1 embattled former 1
## 2 embattled nissan 1
## 3 embattled motor 1
## 4 former nissan 1
## 5 former motor 1
## 6 nissan motor 4pFreq_s<-pFreq[pFreq$Freq>=6,]igraphを利用した描画（無向グラフ）
wng<-as.undirected(graph.data.frame(pFreq_s))
plot(wng)
d3Networkのインストール
install.packages("d3Network")d3SimpleNetwork関数
d3SimpleNetwork(net, width=800, height=500,
standAlone=FALSE, opacity = 1.0,
charge = -200, fontsize = as.numeric(input$sel_fontsize),
parentElement = "#d3networkPlot")netwkアプリケーションの起動(d3Networkを利用)
library(shiny)
runApp("app_netwk")Twitterアプリケーションの作成＆登録
OAuth 認証用
ROAuth, twitteRのインストール
install.packages('twitteR')
install.packages('ROAuth')ROAuth, twitteRのインストールの読み込み
library(twitteR)
library(ROAuth)twitteRからのOauth認証
cacert.pemをダウンロード
download.file(url="http://curl.haxx.se/ca/cacert.pem", destfile="cacert.pem")認証情報(Twitterアプリケーション)
consr_key="***********"
consr_secrt="***********"
req_url ="https://api.twitter.com/oauth/request_token"
acs_url = "https://api.twitter.com/oauth/access_token"
auth_url="https://api.twitter.com/oauth/authorize"cred<-OAuthFactory$new(consumerKey=consr_key,consumerSecret=consr_secrt,requestURL =req_url,accessURL = acs_url,authURL=auth_url)handshake: twitterクライアント接続
cred$handshake(cainfo="cacert.pem")認証情報取得
setup_twitter_oauth(consr_key, consr_secrt, acs_token, acs_token_sec)検索例
searchTwitter("#DH2019", n=10)
searchTwitter("#阪大", n=10)
searchTwitter("#Gohsn", n=10)
searchTwitter("Gohsn + Nissan", n=10)
username = "casualconc"
userTimeline(username)詳細情報取得
stweets<-searchTwitter("#ゴーン", n=10)
stweets[1]
#converting a list type into a dataframe
stweet.df<-twListToDF(stweets)
#colnames
colnames(stweet.df)
cDateTokens<-stweet.df$created
cDateTokens
#cDateFreq<-table(as.Date(cDateLst))
cDateFreq<-table(as.Date(cDateTokens))
cDateTypes<- as.Date(names(cDateFreq))
cDateTokens
cDateFreq
cDateTypesタイムライン情報
tweets<-searchTwitter("#ゴーン", n=100)
#converting a list type into a dataframe
tweet.df<-twListToDF(tweets)
cDateTokens<-tweet.df$created
cDateFreq<-table(as.Date(cDateTokens))
cDateTypes<- as.Date(names(cDateFreq))
cDateFreq
cDateTypesタイムライン表示1
#par(xaxt="s")
par(family="HiraMaruProN-W4")
plot(cDateTypes , as.numeric(cDateFreq), type="h", ylab="frequency")タイムライン表示2
par(xaxt="n")
plot(cDateTypes , as.numeric(cDateFreq), type="h", ylab="frequency")
par(xaxt = "s")
axis.Date(1, las=2,at = seq(min(cDateTypes),max(cDateTypes),"days"),format = "%m/%d")タイムライン表示3
par(xaxt="n")
plot(cDateTypes , as.numeric(cDateFreq), type="h", ylab="frequency")
par(xaxt = "s")
axis.Date(1, las=2,at = seq(min(cDateTypes),max(cDateTypes),"days"),format = "%m月%d日")トレンド情報取得
aTL <- availableTrendLocations()
head(aTL)
subset(aTL, subset=country=="Japan")
subset(aTL, subset=name=="Osaka")
getTrends(subset(aTL, subset=name=="Osaka")[3])