textprepr-vignette

Load package

library(textprepr)

Create tweets data

tweets <- c("Make America Great Again! @DonaldTrump #America",
            "It's rocket-science tier investment~~ #LoveElonMusk",
            "America America America always GREAT",
            "make america great again! #AMERICA")

Data cleaning

remove_punct(tweets)
#> [1] "Make America Great Again DonaldTrump America"  
#> [2] "Its rocketscience tier investment LoveElonMusk"
#> [3] "America America America always GREAT"          
#> [4] "make america great again AMERICA"

remove_punct(tweets, skip="@#'")
#> [1] "Make America Great Again @DonaldTrump #America"  
#> [2] "It's rocketscience tier investment #LoveElonMusk"
#> [3] "America America America always GREAT"            
#> [4] "make america great again #AMERICA"

Feature engineering

extract_ngram(tweets, n=4)
#>  [1] "Make America Great Again!"                     
#>  [2] "America Great Again! @DonaldTrump"             
#>  [3] "Great Again! @DonaldTrump #America"            
#>  [4] "Again! @DonaldTrump #America It's"             
#>  [5] "@DonaldTrump #America It's rocket-science"     
#>  [6] "#America It's rocket-science tier"             
#>  [7] "It's rocket-science tier investment~~"         
#>  [8] "rocket-science tier investment~~ #LoveElonMusk"
#>  [9] "tier investment~~ #LoveElonMusk America"       
#> [10] "investment~~ #LoveElonMusk America America"    
#> [11] "#LoveElonMusk America America America"         
#> [12] "America America America always"                
#> [13] "America America always GREAT"                  
#> [14] "America always GREAT make"                     
#> [15] "always GREAT make america"                     
#> [16] "GREAT make america great"                      
#> [17] "make america great again!"                     
#> [18] "america great again! #AMERICA"

extract_hashtags(tweets)
#> [1] "America"      "LoveElonMusk" "AMERICA"

Data visualization

tweet_count <- generate_cloud(tweets, type="words")

tweet_count <- generate_cloud(tweets, type="hashtag")

tweet_count <- generate_cloud(tweets, type="stopwords")