Today I downloaded tweets for the show #XFactor as this hastag was sort of trending. I didn't want to get one more sports related hashtag. I wasn't too hopeful about this because I thought the hashtag wasn't necessarily trending. But I was wrong! I could quickly get 100K tweets and to my surprise I found that the wordcloud was dominated by certain words ("factor" was sort of obvious but I didn't anticipate it). This is an indication of a lot of retweets. So for me now is the time to adjust the code to remove retweets. But I am not sure whether to remove them because that will mean that we put the same weight on all the tweets not taking into account the popularity (through retweets in this case. I haven't looked into the favorite tweets so far). Anyway, here is the wordcoud.
The word "Olly" shows up because it's the name of the presenter of the show and it seems that he screwed up by telling a contestant the outcome before the results were formally announced.
http://www.telegraph.co.uk/culture/tvandradio/x-factor/11989313/What-time-and-when-is-The-X-Factor-on-TV-this-weekend-Plus-One-Direction-will-perform.html
To my utter surprise and also supporting the view that most look like retweets, the sentiment score was super positive. Also note that the sentiment was available for almost 80% of the tweets, which is very high compared to the past sentiment graphs where this was only about 50%. This all points to a large number of retweets. Maybe someone would like to study who retweeted them. Email me if you want to work on it.
The word "Olly" shows up because it's the name of the presenter of the show and it seems that he screwed up by telling a contestant the outcome before the results were formally announced.
http://www.telegraph.co.uk/culture/tvandradio/x-factor/11989313/What-time-and-when-is-The-X-Factor-on-TV-this-weekend-Plus-One-Direction-will-perform.html
To my utter surprise and also supporting the view that most look like retweets, the sentiment score was super positive. Also note that the sentiment was available for almost 80% of the tweets, which is very high compared to the past sentiment graphs where this was only about 50%. This all points to a large number of retweets. Maybe someone would like to study who retweeted them. Email me if you want to work on it.
Comments
Post a Comment