People had mixed opinions for Rangoon

The recently released Bollywood movie, Rangoon, has been a major topic of discussion this week. It is one of those works of art that get extremely polarized reviews with big chunks of the audience either loving the movie or hating it. We saw this as a perfect playground to deploy our AI algorithms and make sense of the situation. Based on the analysis of more than ten thousand tweets (excluding re-tweets) since the movie was released, we have derived some interesting insights.

Overall, people have mixed opinions

rangoon review
People had mixed opinions for Rangoon

From all the tweets where Rangoon was mentioned, we considered only the ones where the underlying intent was an opinion (28%). We noticed people had mixed feelings about Rangoon compared to clear consensus for movies like Dangal (a super-hit) or Bombay Velvet (considered a flop).

To ensure that our conclusions are legitimate and not noise from probabilistic machine learning models, we compared our results with people’s votes on popular movie portals like IMDb which gives a critic’s perspective and BookMyShow which gives a layman’s perspective. We are pleasantly surprised to see that our AI based ratings (derived from unstructured text in tweets) correlate well with structured user polls.

rangoon review
People had mixed opinions for Rangoon

Direction was the weaker link, Shahid stole the show

We dug deeper to understand how people perceived the basic building blocks of the movie i.e. direction and performance by top stars. In recent years, Shahid Kapoor’s acting has jumped to an altogether different league and Rangoon was no exception. The direction was one aspect that didn’t live up to people’s expectations. People flock to Vishal Bharadwaj’s movies with hopes of a repeat of masterpieces like Omkara and Haider. We concede that its very difficult to consistently live up to those lofty expectations. People seem to have liked the first half (68% positive sentiment) more than the second half (50%).

Kangana’s weaker sentiment score came as a surprise to us, but on digging deeper we found this story as the underlying reason. Kangana had said she was ‘sad’ as many of her favorite scenes were cut by the director. Many people shared this story on Twitter and our algorithm accounted for them as negative sentiment tweets. Deriving such an insight is possible if negative sentiment tweets are analysed in conjunction with the word cloud.

rangoon review
People had mixed opinions for Rangoon

How did they actually describe the movie?

Now comes the interesting part, we take this opportunity to unveil a new feature of Karna-AI where we look at a sentence and identify noun-adjective/adverb pairs. This helped us to deduce how people actually described the movie and performance by key people. Doing something like this manually on a corpus of thousands of tweets is a daunting task, but technology as always saves the day. Take a look at the adjective word cloud associated with each of these entities.

rangoon review
People had mixed opinions for Rangoon

Rave reviews for Shahid Kapoor can be seen in the graphic above whereas for Kangana the word ‘sad’ prominently appears due to the stories about her scenes being cut.

AI is now smart enough for understanding text

At Karna-AI, a ParallelDots product, we deploy a range of cutting-edge AI techniques to derive meaningful insights from large chunks of unstructured text. This can be of great use to brands, market research companies and consulting agencies of all kinds. If you have any questions or suggestions, we would love to hear from you.

Want to get deep consumer insights. Click here to schedule a free demo.

market research

Leave a Reply

Your email address will not be published. Required fields are marked *