Objective:

How to analyse comments from various sources on airlines (and their competitors) to track sentiments associated with them.

Key Challenges:

• Presence of various sources of unstructured data:
  • Customer Reviews
  • Social Media Posts
• Lack of a comprehensive scoring method, combining all the above sources of information

Approach:

• Corpus of vectors (words) made from raw and unstructured data
• Pre-processing engine for white-space removal, punctuation-removal, stop-words removal, etc.
• Term document matrix creation
• Text Classification and NLP Algorithms to categorize each comment into classes
• Sentiment Classifier engine using an ensemble of algorithms to arrive at a weighted sentiment score with optimal weights to reduce errors

Benefits:

• Provided comprehensive customer sentiment score to airline
• Enabled the airline to identify avenues that customers are consistently dissatisfied with
• Provided sentiment scores across time for airlines to correlate the impact of various marketing, promotional and brand-awareness campaigns on the customer sentiments.