A Survey on Box-Office Opening Weekend Prediction Using Twitter Data

Abstract

We live in times where about 1,986 feature films are produced annually and this
amounts to the box office being worth $41.7 billion, making the film industries one of the
biggest markets in the world. In out paper we are only concerned with films produced in
the United States of America, which produces nearly 600 movies annually. We aim to find
out the “buzz” around the movies and therefore predict the opening weekend for each
movie. The opening weekend ticket sales depend upon the buzz the movie has generated
since the announcement, trailers, TV spots and various other factors.Now, the impact of
these trailers, TV spots, interviews on the general audience can be seen in terms of twitter
reactions. This buzz in fact may be negative leading to a lower opening weekend. The buzz
of the movie is the number of people talking about the movie. More the number of people
talking about the movie, greater is the buzz surrounding the movie. To calculate the
number of people talking about the movie we take twitter data into account. Now, people
could be talking positively about the movie or negatively about the movie and we can use
sentimental analysis algorithms to predict that. Finally based on these factors we try to
form a model with best fits our data in order to make future predictions possible.