Extracting the interesting information from the chatter around a TV show is hard - shows a TV show can get 1,000 tweets a minute - nobody can read that many tweets while watching a show.
So we’ve been looking at interesting ways to prioritise and extract data from a show - we’ve built a TweetRank algorithm that in real time rates tweets.
So how good is it?
Let’s compare a wordle of all the tweets about last night’s Four Rooms on Channel 4
With a wordle of the 500 most interesting tweets - which we hope, if TweetRank is doing it’s job right, and extracting the most interesting tweets, will look similar - even though it’s operating on a much smaller data set than the whole stream.
They look pretty similar to us - so we are happy that the output of the TweetRank algorithm gives viewers a reasonable picture of all the tweets about a show without having to read all of them
Discussion
No comments yet.