Powered by ScribbleLive
The Red Carpet Rumble was built to analyze Twitter reaction to celebrities on the red carpet of the Oscars. We are working with Pointslocal to evaluate the massive amounts of chatter on Twitter to determine which celebrity wore it best (or worst) during one of the most anticipated Hollywood events of the year. Many people are watching ... and Twitter is on fire. The celeb with the top social score, based on the formula explained below, will be named the winner of Red Carpet Rumble 2013.
First we ingest the total number of tweets about a celebrity based on selected keywords. We are able to load the system with a number of keywords prior to the festivities. Celebrity names can have a number of different variations (J-Lo, Brangelina, Bennifer, for instance). But then we will add on the fly, allowing for a wider scope to ensure as many tweets about a celebrity are calculated. Especially if something bizarre or unexpected happens.
Second, we use a natural language processor to determine the opinion of the tweet (whether it's positive or negative). Yes, we even have a sarcasm filter. Whether the tweet is positive or negative will provide a "net sentiment score." The net sentiment score is defined by the formula:
(Positive tweets + Neutral tweets - Negative tweets)/ Total celebrity tweets.
Finally, to rank the celebs, we calculate a "Red Carpet Rumble" score, in which we divide by the total tweets for all celebrities:
(Positive tweets + Neutral tweets - Negative tweets)/ Total tweets for all celebrities.
The scoreboard will update as close to real time as possible. The window for scanning tweets will open at 4:30 EST (one hour prior to the live broadcast of the red carpet) and will end at the start of the Oscars. This will prevent any pre-carpet or Oscars show hype to influence the overall scores. Once the window closes, we will do a final calculation and then announce the overall winner, the most talked about, and most loved.