Use Social Media Chatter to Predict the Success of a Movie in Television​

The client is a Global Media and Entertainment conglomerate with a strong presence across television broadcasting, movies, music, live entertainment and digital businesses. Spread over 173 countries, connecting and entertaining over 1.3 billion people globally

Business Challenges

  • The client buys world TV premier rights for movies 
  • These movies were already released in theatres 
  • The success of a movie in the theatre did not guarantee a high viewership on TV 

Our Solution

We looked at social media data on the movie as a proxy for viewer engagement. Based on the social media Centelon built an AI based model to predict the viewership on Television. The solution included the following

  • Automatic data cleaning as the data included images, special characters, and slang 
  • Data visualization on engagement based on their gender, location, age 
  • Movie engagement in terms of genre of the movie, music, trailer 
  • Sentiment model that would include Hindi and English 
  • Predictive model for viewership 
  • Attribution of various features in the predictive model 


  • The model was tested on new data and it predicted the viewership with high accuracy 
  • The client was able to make better decision the commercial decision of buying movie rights based on the model.