Can We Predict Rotten Tomatoes Ratings?When is a movie actually rotten or fresh?Esther L.BlockedUnblockFollowFollowingDec 5Movie review sites such as IMDb and Rotten Tomatoes have become a part of many movie goers’ lives.Movie watchers can easily determine what movie to watch and compare certain movies by looking at a single number — a movie rating between 0% and 100%.But… have you ever wondered how those ratings are determined?Picture Source: http://you-deserve-this.blogspot.com/2012/08/imdb-top-250-vs-rotten-tomatoes.htmlIt is not rare that some box-office hits and “classics” have rather low movie ratings on these websites..This has led us to wonder if the movie reviews that you see online actually reflect the general public’s opinions?When looking at IMDb and Rotten Tomatoes, we see that both movie aggregator sites do not state how their movie rating percentages are calculated.IMDb’s Help page:We take all the individual ratings cast by IMDb registered users and use them to calculate a single rating..As a result, we worked with Rotten Tomatoes for the purposes of our project.Building a Movie Scoring ModelWe were most interested in answering two key questions.Does Rotten Tomatoes emphasize certain aspects of movie reviews more than others in its calculation of its Audience Rating?Is there a relationship between Rotten Tomatoes’ Audience Rating and its Tomatometer rating?During our quest to see if we could create our own scoring mechanism, there were several steps we took.Stage 1: Defining the DatasetWe manually selected about 150 movies that had Rotten Tomatoes Audience Scores ranging from 0–100%..For consistency, we will refer to movies ranging from 0–25% as the first quartile, 26–50% as the second, 51–75% as the third, and finally 76–100% as the fourth quartile.Distributions Across Tomatometer and Audience Ratings in Dataset: ObservationsOne thing that we noticed while building our list of movies was that movies on Rotten Tomatoes rarely have an audience rating below 25%..As a result, you can see that while the number of movies with ratings in the upper three quartiles is relatively uniform, there is a notable disparity between the movie counts in the upper quartiles and that of the first quartile.We believe this lack of movies with very low audience ratings is due to the fact that audience ratings are being produced by a large quantity of Rotten Tomatoes users that have no authority or accreditation to rate movies..Or possibly, because these ratings are produced from all users, they more accurately portray the general publics opinion and people typically tend to like most movies and give them a higher rating.Number of Movies for each Quartile for Tomatometer and Audience ScoresInterestingly enough, although our dataset was built by Audience Score, we notice a lower degree of variance across quartile counts for Tomatometer scores.We believe that this is because the Tomatometer rating is produced from legitimate, accredited movie and TV critics..Our classifier has shown us that using the audience reviews on Rotten Tomatoes, we were not consistently able to accurately predict the audience rating nor Tomatometer rating of these movies..So clearly, there are other factors being considered when determining the ratings displayed on the Rotten Tomatoes website.SummaryFrom our study, we are able to determine that the movie rating scores on Rotten Tomatoes, and most likely on other websites as well, are not accurately displaying the general public’s opinion on the movie..Some notable words from audience movie reviews are shown in the word cloud below:Word Cloud Generated from Review Text InputWe advise movie watchers to be wary of the movie ratings they see displayed on movie websites such as Rotten Tomatoes and IMDb.. More details