Social Opinion Mining Platform

A cloud-based platform that identifies different forms of opinions, such as objectivity, subjectivity, sentiment, emotion and sarcasm, from social data (mostly semi-structured and unstructured) extracted across different sources, such as digital newspapers, blogs, social networks (e.g., Facebook) and microblogs (e.g., Twitter) and represented in various media formats (text, image, audio, video).

The Social Opinion Mining components are based on natural language processing, machine learning and deep learning techniques. They can be adapted for multiple domains, such as Finance, Politics, Government, and iGaming, that have the potential to make a real-world impact.

The platform is built as a set of microservices, whereby each microservice exposes a different opinion mining technique. The advantages of this architecture is that it is possible to scale each service independently, according to needs. It is also possible to use a mix of technologies, which allows for optimisation per microservice.


  • Sentiment: classifies the polarity of an online post and identifies its intensity
  • Emotion: identifies the human emotional states expressed in an online post
  • Sarcasm: determines whether an online post is hurtful to the person being said
  • Irony: determines whether an online post conveys the opposite of what is being said