Twitter

Twitter Sentiment Analysis Data-set

This is an entity-level sentiment analysis data-set of twitter. Given a message and an entity, the task is to judge the sentiment of the message about the entity. There are three classes in this data-set: Positive, Negative and Neutral. We regard messages that are not relevant to the entity (i.e. Irrelevant) as Neutral.

Dimension of the data-set

[1] 74682     4

Variables in the data-set

[1] "ID"        "Topic"     "Sentiment" "Text"     

Description of the variables in the dataset

'data.frame':   74682 obs. of  4 variables:
 $ ID       : int  2401 2401 2401 2401 2401 2401 2402 2402 2402 2402 ...
 $ Topic    : chr  "Borderlands" "Borderlands" "Borderlands" "Borderlands" ...
 $ Sentiment: chr  "Positive" "Positive" "Positive" "Positive" ...
 $ Text     : chr  "im getting on borderlands and i will murder you all ," "I am coming to the borders and I will kill you all," "im getting on borderlands and i will kill you all," "im coming on borderlands and i will murder you all," ...

Summary of the data-set

       ID           Topic            Sentiment             Text          
 Min.   :    1   Length:74682       Length:74682       Length:74682      
 1st Qu.: 3195   Class :character   Class :character   Class :character  
 Median : 6422   Mode  :character   Mode  :character   Mode  :character  
 Mean   : 6433                                                           
 3rd Qu.: 9601                                                           
 Max.   :13200                                                           

Checking for the missing values

[1] 0

Column diagram for differnt topics vs count

Donut chart for Sentiment (Positive,Neutral,Negative,Irrelevant)

Bar chart for different levels of sentiment (Positive, Neutral, Negative, Irrelevant) vs. different topics

Warning in geom_bar(aes(fill = Sentiment), stat = "identity", positive =
"dodge"): Ignoring unknown parameters: `positive`

We list the Twitter topics in decreasing order of their trending popularity.

                        MaddenNFL                         Microsoft 
                             2400                              2400 
             TomClancysRainbowSix                        CallOfDuty 
                             2400                              2394 
                  LeagueOfLegends                           Verizon 
                             2394                              2382 
                      ApexLegends         CallOfDutyBlackopsColdWar 
                             2376                              2376 
                         Facebook                             Dota2 
                             2370                              2364 
                     WorldOfCraft                             NBA2K 
                             2364                              2352 
                      Battlefield              TomClancysGhostRecon 
                             2346                              2346 
                             FIFA                         Overwatch 
                             2340                              2334 
                    Xbox(Xseries)                   johnson&johnson 
                             2334                              2328 
                           Amazon                         HomeDepot 
                             2316                              2310 
                PlayStation5(PS5)                             CS-GO 
                             2310                              2304 
                    Cyberpunk2077               GrandTheftAuto(GTA) 
                             2304                              2304 
                           Google                       Hearthstone 
                             2298                              2298 
                           Nvidia                       Borderlands 
                             2298                              2286 
                         Fortnite PlayerUnknownsBattlegrounds(PUBG) 
                             2274                              2274 
           RedDeadRedemption(RDR)                    AssassinsCreed 
                             2262                              2244 

Sentiment Distribution for ‘Google’,‘Facebook’,‘Microsoft’ in Twitter

Histogram of Text Length in Tweets

This histogram visualizes the distribution of tweet lengths based on the number of characters.

The distribution is right-skewed, with the majority of tweets having shorter text lengths.

The highest frequency of tweets falls within the 0 to 100 character range, with over 20,000 tweets in this interval. This indicates that most tweets are concise.

As text length increases, the frequency of tweets decreases sharply. Very few tweets exceed 300 characters, and tweets with lengths approaching the maximum of 1,000 characters are extremely rare.

This heatmap visualizes the relationship between sentiment (Positive, Neutral, Negative, Irrelevant) and various topics discussed on Twitter. Each cell in the heatmap represents the intensity or frequency of a particular sentiment associated with a specific topic, allowing for a quick identification of patterns and trends across different topics.

Word Cloud

In a word cloud, the size of each word indicates its frequency or importance—the larger the word, the more frequently it appears in the text.

Loading required package: RColorBrewer

Loading required package: NLP

Attaching package: 'NLP'

The following object is masked from 'package:ggplot2':

    annotate

Dominant Words: The largest words like “game,” “just,” “like,” “will,” and “good” are the most frequently mentioned in the dataset. This suggests that the tweets may be heavily focused on gaming-related discussions.

Sentiment and Topics: Words like “good,” “love,” and “great” suggest positive sentiment, while words like “fix,” “shit,” and “fucking” might indicate negative sentiment or frustration. The word “game” is central, which could imply that the primary topic of discussion is gaming.

Trends: The variety of words related to gaming, companies (e.g., “Verizon,” “Google,” “Amazon”), and social media engagement (e.g., “Facebook,” “Twitter”) indicate the topics that are trending or commonly discussed in the dataset.