Lec11: Twitter API v2

Lec11

Twitter APIの利用

References

• Developer Platform
• Developer Platform: Docs
• Introduction to Twitter API v2
• Francesco Bailo. 2022. “Twitter API Academic Research access with R”
• RTwitterV2

Install packages: httr, jsonlite

install.packages("httr")
install.packages("jsonlite")
#install.packages("dplyr")
install.packages("plyr")

パッケージの読み込み

library(httr)
library(jsonlite)
library(dplyr)
library(plyr)

Bearer Authentication

headers <- c(`Authorization` = sprintf('Bearer %s',  Sys.getenv("TWITTER_BEARER")))

Get tweets with specific words

RecentSearchURL <- 'https://api.twitter.com/2/tweets/search/recent'
params <- list(query = "(chatGPT OR OpenAI)(-is:retweet)", tweet.fields = "author_id,created_at,public_metrics", max_results = 20)
res <- httr::GET(url = RecentSearchURL, httr::add_headers(.headers = headers), query = params)

res

Review the response in the JSON format

obj <- httr::content(res, as = "text")
#prettify(obj)

Review the response in the JSON format

• pipes: “%>%”

tweetData <- fromJSON(obj, flatten = TRUE) %>% as.data.frame()
#tmp <- fromJSON(obj, flatten = TRUE) 
#as.data.frame(tmp)

Review the response in a tibble

as_tibble(head(tweetData))

Get tweets with specific words & language

Specify English langage: “lang:en”

params <- list(query = "(chatGPT OR OpenAI)(lang:en)(-is:retweet)", tweet.fields = "author_id,created_at,public_metrics", max_results = 20)
res <- httr::GET(url = RecentSearchURL, httr::add_headers(.headers = headers), query = params)

Review the response

obj <- httr::content(res, as = "text")
tweetData <- fromJSON(obj, flatten = TRUE) %>% as.data.frame()
as_tibble(head(tweetData))

Specify English langage: “lang:ja”

params <- list(query = "(chatGPT OR OpenAI)(lang:ja)(-is:retweet)", tweet.fields = "author_id,created_at,public_metrics", max_results = 20)
res <- httr::GET(url = RecentSearchURL, httr::add_headers(.headers = headers), query = params)

Review the response

obj <- httr::content(res, as = "text")
tweetData <- fromJSON(obj, flatten = TRUE) %>% as.data.frame()
as_tibble(head(tweetData))

Remove “t.co” links in the “data.text” column

tweetText <- gsub("http\\S+", "", tweetData$data.text)

Export the data

write(tweetText, "recent_tweets_ja.txt")

日本語テキスト処理

RMeCabの読み込み

path_home <- system("echo $HOME",intern=T)
lib_path <- paste(path_home,"/usr/local/lib/libmecab.so.2", sep="")
dyn.load(lib_path)
library(RMeCab)

形態素解析

tweetsDF <- docDF("recent_tweets_ja.txt", type=1)

file_name =  ./recent_tweets_ja.txt opened
number of extracted terms = 282
now making a data frame. wait a while!

sub_tweetsDF <- tweetsDF %>% filter(POS1 %in% c("名詞", "形容詞", "副詞"), POS2 %in% c("一般", "固有名詞","自立"))
head(sub_tweetsDF)

Wordcloud visualization

library(wordcloud2)
sub_tweetsDF[,c(1,4)]

wordcloud2(sub_tweetsDF[,c(1,4)], size = 0.8, shape = 'pentagon')

Specify pipes: “%>datetime values

params <- list(query = "(chatGPT OR OpenAI)(lang:en)(-is:retweet)", tweet.fields = "author_id,created_at,public_metrics", start_time = "2023-01-08T00:00:00Z", end_time = "2023-01-09T00:00:00Z", max_results = 20)

httr::GET(url = RecentSearchURL, httr::add_headers(.headers = headers), query = params) %>% httr::content(., as = "text") %>% fromJSON(., flatten = TRUE)  %>% as.data.frame() -> tweetData2 
as_tibble(head(tweetData2))

params <- list(query = "(chatGPT OR OpenAI)(lang:en)(-is:retweet)", tweet.fields = "author_id,created_at,public_metrics", start_time = "2023-01-10T00:00:00Z", end_time = "2023-01-12T00:00:00Z", max_results = 20)

httr::GET(url = RecentSearchURL, httr::add_headers(.headers = headers), query = params) %>% httr::content(., as = "text") %>% fromJSON(., flatten = TRUE)  %>% as.data.frame() -> tweetData3 
as_tibble(head(tweetData3))

library("wordcloud")

Loading required package: RColorBrewer

#colnames(tweetData2)
gsub("http\\S+", "", tweetData2$data.text) %>% gsub("(\\n|&gt)", "", .) %>% strsplit(.,"[[:space:]]|[[:punct:]]") %>% unlist %>% .[. != ""]  %>% table -> wordLst2

wordcloud(names(wordLst2), wordLst2)

感情分析

• Julia Silge & David Robinson (2017) “Text Mining with R”, O’Reilly
• 日本語評価極性辞書