Topic classification and sentiment analysis for a top retailer
- Content from customer reviews, surveys, social media posts, etc. were leveraged.
- Our text classification engine (that leverages various supervised learning algorithms) was used to categorize each comment into classes such as customer service, pricing, in-store experience, etc.
- Our sentiment analysis engine used an ensemble of algorithms to arrive at a weighted sentiment score for each comment within each of the classes of comments.
- The solution was built leveraging tcg mcube, our proprietary analytics platform that can analyse large volumes of unstructured data to drive accelerated insights. The built-in data models and data science algorithms drive velocity to value for our customers.
- Revised brand strategy helped ameliorate customer sentiments, reduced churn by around 8% and increased revenue by 5%.