Twitter and the Prediction of Oscar Winners

Predicting winners of the Oscars through data has gained increasing interest in recent years. There have been many efforts to predict Oscar winners, but one important data element has been excluded from the analyses, microblogging sites (MS) and their effect on public opinion. Therefore, we analyze the correlation of MS activity and Oscar winners from 2015 to 2019 through retweets, mentions, hashtags, and favorites on Twitter. This analysis is performed through statistical analyses, logistic regression, and time series decomposition to determine whether this factor could be related to Oscar winner forecasting. Some significant findings were that best picture winners had significantly less total retweets (393 (IQR 367.5-415)) than non-winning nominees (548 (IQR 431.5-989)), P=0.016. In the future, it may be interesting to add years and categories to the analysis and to study Twitter results by geographic location or online blogs dedicated to movie critics and reviews.