Given a list of competing products (on Amazon) and relevant features, this algorithm will perform feature-specific sentiment analysis on each product and each feature. This is useful to 1) understand what consumers like/dislike about your product and your competitors, 2) identify market opportunities through pain-points, and 3) estimate the "shadow price" of each unit on the willingness to pay of your customers.
>> Blog post with demo coming soon.
Input, object with two properties:
- products: array of objects where each object represents a product with a name (string) and url (string) attributes
- features: array of features to be examined
Output: array of objects, with each object having a name (string), url (string), and features (object). A Feature object has a name (string), sentences (string), scores (double), avg_score (double).