Venga Sentiment Analysis with IBM's Watson
Venga measures the sentiment behind a reviewers comments by analyzing each review word by word. It takes a lot of work but fortunately Venga has IBM’s Watson as a resource. By passing the review text through Watson’s AI, Venga is able to determine the sentiment of each word or phrase in a review, and assign it to the appropriate sentiment category (Services, Staff, Facilities, Equipment, Value). The result is the detailed analysis on the Sentiment page of the Feedback tab. Here's how it works:
New Reviews are Analyzed and Scored
- When a review is submitted to Venga, it is sent to Watson for analysis.
- Watson’s AI reads the review and identifies Key Terms. A key term is any word or phrase that contributes to the sentiment of the review. This includes references to the studio’s services, staff, facilities, equipment, or value.
- Watson’s AI determines the sentiment of each key term by examining the context around it.
- Key terms are assigned a numerical score. For example, if a review mentions the studio’s locker room, Watson looks for specific words related to the locker room that carry positive or negative sentiment to determine how the reviewer felt about the dish. Positive context like clean, or comfortable give the locker room a high score. Dirty and smelly carry a highly negative sentiment and result in a low score. More neutral words like average, or ok, are scored in the middle.
Key Terms are Categorized
- After each key term has been scored, the review is returned to Venga for categorization.
- Each key term is associated with one of the five sentiment categories, Services, Staff, Facilities, Equipment, Value
Sentiment Categories Receive a Letter Grade
- Venga calculates a score for each category by averaging Watson's sentiment scores for the words in that category. The Overall score is an average of the scores for the individual categories.
- The score is converted to a letter grade scale. An A+ is great and a D means there is work to be done.
- Click a letter grade to Investigate the key terms associated with it. Clicking a letter grade opens a window displaying a grade trend graph and a list of the keywords that contributed to the letter grade.
Learn More: Feedback Dashboard: Tips and Best Practices