Product review sentiment analysis with a summary in chat

Every ten minutes the scenario pulls fresh reviews, judges the tone of each, files them into a table and posts a summary to chat: how many positive, negative and neutral.

  • A bad review surfaces the same hour
  • Every review is stored with its tone and product
  • The summary shows the whole picture, not isolated shouts
  • Batch processing keeps big review volumes from breaking the run