Corpus with Book Reviews from Bokelskere.no

This corpus is a dump of user generated book reviews and discussions from Bokelskere.no (meaning "book lovers"), a web community where users review and discuss new and old literature, both fiction and non-fiction. The corpus is structured as a JSON Array where each object corresponds to a review or comment to a review on Bokelskere.no. Each object has the following fields: - "post_id": unique identifier for review - "date": date when the review was posted - "user_id": unique identifier for the user - "isbn13": ISBN for the the rewieved book - "post_title": title of review - "text": review - "score": evaluation (from 1-6, where 6 is the best) - "main_title": title of reviewed book - "author": author of reviewed book - "parent_id": identifier of review which has been commented upon The corpus contains approximately 219,000 posts/objects, and 1.5 million word tokens (in the "text"-field).